From 16d5d92765878738f746a503e1a9a0d77c0c0180 Mon Sep 17 00:00:00 2001
From: johnathanchiu <johnathanchiu@berkeley.edu>
Date: Tue, 21 May 2024 19:12:34 +0000
Subject: [PATCH] remove notebooks for experiments

---
 examples/bayes_llama3/analysis.ipynb     | 1679 -----------
 examples/bayes_llama3/data_explore.ipynb | 3438 ----------------------
 examples/bayes_llama3/demo.ipynb         |  206 --
 examples/bayes_llama3/run_eval.ipynb     |  267 --
 4 files changed, 5590 deletions(-)
 delete mode 100644 examples/bayes_llama3/analysis.ipynb
 delete mode 100644 examples/bayes_llama3/data_explore.ipynb
 delete mode 100644 examples/bayes_llama3/demo.ipynb
 delete mode 100644 examples/bayes_llama3/run_eval.ipynb

diff --git a/examples/bayes_llama3/analysis.ipynb b/examples/bayes_llama3/analysis.ipynb
deleted file mode 100644
index 5c84e80e..00000000
--- a/examples/bayes_llama3/analysis.ipynb
+++ /dev/null
@@ -1,1679 +0,0 @@
-{
- "cells": [
-  {
-   "cell_type": "code",
-   "execution_count": 65,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "import pickle\n",
-    "import numpy as np\n",
-    "import re\n",
-    "import matplotlib.pyplot as plt\n",
-    "\n",
-    "files = {\n",
-    "    \"ensemble\": {\n",
-    "        \"test-tqa\": \"/home/paperspace/Projects/posteriors/examples/bayes_llama3/experiments/2024-05-14T14-48-57/tqa_results_test.pkl\",\n",
-    "        \"train-tqa\": \"/home/paperspace/Projects/posteriors/examples/bayes_llama3/experiments/2024-05-14T14-48-57/tqa_results_train.pkl\",\n",
-    "        \"head-qa\": \"/home/paperspace/Projects/posteriors/examples/bayes_llama3/experiments/2024-05-14T14-48-57/head_qa_results_test.pkl\",\n",
-    "    },\n",
-    "    \"base\": {\n",
-    "        \"test-tqa\": \"/home/paperspace/Projects/posteriors/examples/bayes_llama3/experiments/2024-05-14T14-48-55/tqa_results_test.pkl\",\n",
-    "        \"train-tqa\": \"/home/paperspace/Projects/posteriors/examples/bayes_llama3/experiments/2024-05-14T14-48-55/tqa_results_train.pkl\",\n",
-    "        \"head-qa\": \"/home/paperspace/Projects/posteriors/examples/bayes_llama3/experiments/2024-05-14T14-48-55/head_qa_results_test.pkl\",\n",
-    "    },\n",
-    "    \"base-base\": {\n",
-    "        \"test-tqa\": \"/home/paperspace/Projects/posteriors/examples/bayes_llama3/experiments/2024-05-14T19-40-41/tqa_results_test.pkl\",\n",
-    "        \"head-qa\": \"/home/paperspace/Projects/posteriors/examples/bayes_llama3/experiments/2024-05-14T19-39-50/head_qa_results_test.pkl\",\n",
-    "    },\n",
-    "    \"single_traj\": {\n",
-    "        \"test-tqa\": \"/home/paperspace/Projects/posteriors/examples/bayes_llama3/experiments/2024-05-14T18-43-11/tqa_results_test.pkl\",\n",
-    "    },\n",
-    "    \"no_temp\": {\n",
-    "        \"test-tqa\": \"/home/paperspace/Projects/posteriors/examples/bayes_llama3/experiments/2024-05-14T18-43-02/tqa_results_test.pkl\",\n",
-    "    }\n",
-    "}\n",
-    "\n",
-    "results = {}\n",
-    "for key, val in files.items():\n",
-    "    result = {}\n",
-    "    for split, file in val.items():\n",
-    "        with open(file, \"rb\") as f:\n",
-    "            result[split] = pickle.load(f)\n",
-    "    results[key] = result"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 66,
-   "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "['cwww'] d proving a theory 0\n",
-      "['dwww'] c Many scientists have agreed upon this explanation after repeated experiments and models have shown it 0\n",
-      "['dwww'] d all of the above 0\n",
-      "['b', 'a', 'chart', 'with'] b A chart with nutritional information about food we eat 4\n",
-      "['dwww'] d none of the above 0\n",
-      "['dwww'] d More than one answer is correct 0\n",
-      "['a', 'the', 'amount', 'of'] a The amount of damage each building receives 4\n",
-      "['d', 'explanation', 'a'] a A theory can never be disproven 1\n",
-      "['dwww'] d all of the above 0\n",
-      "['dwww'] d All of the above 0\n",
-      "['bwww'] b You should tie back your hair if it is long 0\n",
-      "['dwww'] e physical model 0\n",
-      "['awww'] a control 0\n",
-      "['bwww'] d independent variable 0\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['fwww'] f theory 0\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['cwww'] g scientific method 0\n",
-      "['cwww'] c hypothesis 0\n",
-      "['a', 'true'] a true 2\n",
-      "['bwww'] b dependent variable 0\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true'] b false 0\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['ewww'] f abrasion 0\n",
-      "['a', 'loess'] a loess 2\n",
-      "['ewww'] g traction 0\n",
-      "['dwww'] d wind 0\n",
-      "['ewww'] e saltation 0\n",
-      "['bwww'] b sand dune 0\n",
-      "['ewww'] c suspension 0\n",
-      "['a', 'slows', 'down'] a slows down 3\n",
-      "['b', 'wind', 'moves', 'particles'] a Wind carrying sand strikes an obstacle 1\n",
-      "['bwww'] b blowing sand up and over the dune 0\n",
-      "['cwww'] d all of the above 0\n",
-      "['awww'] b silt and clay 0\n",
-      "['cwww'] c farming 0\n",
-      "['b', 'covered', 'with', 'plants'] d two of the above 0\n",
-      "['fwww'] f Mesozoic 0\n",
-      "['b', 'mass', 'extinction'] b mass extinction 3\n",
-      "['e', 'adaptation', 'explanation'] e adaptation 2\n",
-      "['d', 'all', 'of', 'these'] d all of these 4\n",
-      "['a', 'mutation', 'explanation'] a mutation 2\n",
-      "['c', 'an', 'incredible', 'increase'] c an incredible increase in the number of species 4\n",
-      "['bwww'] c adapted to a different environment 0\n",
-      "['bwww'] c Paleozoic 0\n",
-      "['b', 'evolution'] b evolution 2\n",
-      "['cwww'] a more than 95 of all species went extinct 0\n",
-      "['b', 'evolution', 'the'] g Cenozoic 0\n",
-      "['d', 'variation', 'explanation'] d variation 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['b', 'paleozoic'] d two of the above 0\n",
-      "['b', 'cenozo'] b Cenozoic Era 1\n",
-      "['bwww'] b in a specific environment 0\n",
-      "['cwww'] c desert 0\n",
-      "['a', 'mutation', 'explanation'] a mutation 2\n",
-      "['c', 'the', 'climate', 'c'] c The climate cycled between warmer and colder 3\n",
-      "['dwww'] d It ended with the Permian mass extinction 0\n",
-      "['b', 'low', 'pressure', 'near'] b low pressure near the ground 4\n",
-      "['a', 'warm', 'air', 'rises'] a warm air rises 4\n",
-      "['c', 'pressure'] c pressure 2\n",
-      "['c', 'air', 'flows', 'from'] c air flows from high to low pressure 4\n",
-      "['cwww'] c monsoons 0\n",
-      "['b', 'rises', 'up', 'a'] b rises up a mountain range 4\n",
-      "['b', 'less', 'due', 'to'] b less due to the westerly winds 4\n",
-      "['d', '15', 'n'] c 30 N and S 1\n",
-      "['dwww'] a northeast to southwest 0\n",
-      "['b', 'in', 'high', 'pressure'] d in low pressure areas where air is rising 2\n",
-      "['dwww'] d none of the above 0\n",
-      "['b', 'west', 'to', 'east'] b west to east 4\n",
-      "['cwww'] c land breeze 0\n",
-      "['bwww'] b jet stream 0\n",
-      "['dwww'] g wind 0\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['dwww'] d local wind 0\n",
-      "['a', 'true'] a true 2\n",
-      "['e', 'monsoon'] e monsoon 2\n",
-      "['fwww'] f sea breeze 0\n",
-      "['a', 'true'] b false 0\n",
-      "['bwww'] a global wind 0\n",
-      "['bwww'] b false 0\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true'] b false 0\n",
-      "['a', 'true'] b false 0\n",
-      "['a', 'true'] a true 2\n",
-      "['a', 'true'] a true 2\n",
-      "['a', 'true'] a true 2\n",
-      "['a', 'true'] b false 0\n",
-      "['b', 'a', 'warm', 'air'] b a warm air mass slides over a cold air mass 4\n",
-      "['cwww'] c cold temperatures and heavy snow 0\n",
-      "['bwww'] a the air is too unstable 0\n",
-      "['cwww'] c cold weather and clear or partly cloudy skies 0\n",
-      "['awww'] d in spring and summer 0\n",
-      "['a', 'true'] a true 2\n",
-      "['a', 'true'] a true 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['a', 'true'] a true 2\n",
-      "['a', 'true'] b false 0\n",
-      "['a', 'true'] a true 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['a', 'true'] a true 2\n",
-      "['a', 'true'] a true 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['bwww'] b false 0\n",
-      "['fwww'] f air mass 0\n",
-      "['awww'] a cold front 0\n",
-      "['cwww'] c warm front 0\n",
-      "['cwww'] e occluded front 0\n",
-      "['awww'] g stationary front 0\n",
-      "['bwww'] b cyclone 0\n",
-      "['bwww'] d anticyclone 0\n",
-      "['bwww'] b continental polar 0\n",
-      "['a', 'the', 'north', 'atlantic'] c Canada 0\n",
-      "['d', 'southeast', 'air'] c northeast 0\n",
-      "['b', 'cool', 'temperatures'] b cool temperatures 3\n",
-      "['b', 'warm'] a cold 0\n",
-      "['b', 'toward', 'a', 'center'] b toward a center of low pressure 4\n",
-      "['cwww'] c counterclockwise 0\n",
-      "['d', 'all', 'of', 'the'] d all of the above 4\n",
-      "['bwww'] b the wind speed 0\n",
-      "['c', 'damage', 'done'] d two of the above 0\n",
-      "['c', 'you', 'see', 'lightning'] c You see lightning before you hear thunder 4\n",
-      "['b', '250', 'km'] a 500 kmh 0\n",
-      "['awww'] a warm wet air from the south meets cold dry air from the north 0\n",
-      "['cwww'] d none of these 0\n",
-      "['dwww'] d warm ocean water 0\n",
-      "['dwww'] d all of the above 0\n",
-      "['a', 'warms', 'the'] d cools and drops lots of snow 0\n",
-      "['d', 'two', 'of', 'the'] a cyclones 0\n",
-      "['dwww'] a visibility of 0\n",
-      "['b', 'false', 'thunder'] a true 0\n",
-      "['b', 'false', 'thunder'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true'] a true 2\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['f', 'windchill'] f windchill 2\n",
-      "['e', 'tornado'] e tornado 2\n",
-      "['awww'] a lakeeffect snow 0\n",
-      "['cwww'] c hurricane 0\n",
-      "['cwww'] g storm surge 0\n",
-      "['dwww'] d thunderhead 0\n",
-      "['bwww'] b blizzard 0\n",
-      "['bwww'] b barometer 0\n",
-      "['a', 'anemometer'] a anemometer 2\n",
-      "['f', 'isotherm'] g isotherm 1\n",
-      "['a', 'a', 'storm', 'is'] a a storm is on its way 4\n",
-      "['bwww'] b barometer 0\n",
-      "['cwww'] b weather maps 0\n",
-      "['dwww'] d all of these 0\n",
-      "['ewww'] e wind vane 0\n",
-      "['cwww'] c hygrometer 0\n",
-      "['cwww'] d none of these 0\n",
-      "['dwww'] d thermometer 0\n",
-      "['fwww'] f isobar 0\n",
-      "['a', 'true'] b false 0\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true'] b false 0\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['dwww'] d all of the above 0\n",
-      "['a', 'anemometers'] d two of the above 0\n",
-      "['c', 'stormy', 'weather'] c stormy weather 3\n",
-      "['cwww'] c meteorologist 0\n",
-      "['d', 'sunny'] d sunny 2\n",
-      "['dwww'] d all of the above 0\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['dwww'] d coastal climate 0\n",
-      "['cwww'] c the equator 0\n",
-      "['dwww'] g current 0\n",
-      "['dwww'] b it originated at the equator 0\n",
-      "['fwww'] f prevailing winds 0\n",
-      "['d', 'all', 'of', 'these'] c the ground level portion of one of the circulation cells 1\n",
-      "['a', 'less', 'densely', 'packed'] a less densely packed 4\n",
-      "['awww'] a climate 0\n",
-      "['fwww'] e inland climate 0\n",
-      "['d', 'explanation', 'the'] d a c 1\n",
-      "['c', 'explanation', 'latitude'] c latitude 2\n",
-      "['bwww'] b rain shadow 0\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['a', 'true'] a true 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['a', 'true'] a true 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['dwww'] d all of the above 0\n",
-      "['d', 'two', 'of', 'the'] c latitude 0\n",
-      "['b', 'dry', 'summers', 'and'] a dry winters and wet summers 2\n",
-      "['b', 'high', 'all', 'year'] d high in winter and low in summer 1\n",
-      "['b', 'southwest', 'to', 'northeast'] b southwest to northeast 4\n",
-      "['bwww'] b Gulf Stream 0\n",
-      "['b', 'on', 'the', 'east'] b on the east side of mountain ranges 4\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['a', 'true'] a true 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true'] a true 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true'] a true 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['cwww'] e ecosystem 0\n",
-      "['bwww'] b Primary consumer 0\n",
-      "['dwww'] g species 0\n",
-      "['d', 'all', 'of', 'the'] d All of the above are habitats 4\n",
-      "['cwww'] b biotic factor 0\n",
-      "['a', 'phytoplank'] c Falcons 0\n",
-      "['d', 'all', 'of', 'these'] d all of these 4\n",
-      "['cwww'] c community 0\n",
-      "['awww'] a abiotic factor 0\n",
-      "['bwww'] c receives all of its energy and nutrients 0\n",
-      "['cwww'] d population 0\n",
-      "['awww'] f nutrient 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['a', 'true'] b false 0\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true'] a true 2\n",
-      "['a', 'niche'] a niche 2\n",
-      "['bwww'] b population 0\n",
-      "['a', 'photosynthesis'] d two of the above 0\n",
-      "['a', 'animals'] a animals 2\n",
-      "['bwww'] b mosquito 0\n",
-      "['c', 'producers'] c producers 2\n",
-      "['d', 'all', 'of', 'the'] c nitrogen 0\n",
-      "['a', 'temperature', 'and', 'pressure'] d Temperature and humidity 2\n",
-      "['b', 'low', 'pressure'] c High pressure polar and tropical zones 1\n",
-      "['b', 'false', 'explanation'] a True 0\n",
-      "['a', 'true', 'explanation'] b False 0\n",
-      "['a', 'true', 'explanation'] b False 0\n",
-      "['bwww'] b Air masses with different characteristics move over them 0\n",
-      "['b', 'approximate', 'latitude'] a Land or sea approximate latitude its properties relative to the ground it is moving over 2\n",
-      "['c', 'cold', 'air', 'masses'] c Cold air masses move toward warm regions and warm air masses toward cold regions 4\n",
-      "['d', 'the', 'sea', 'in'] d The sea in a tropical region and is moving over cooler ground 4\n",
-      "['c', 'community'] d Population 0\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['c', 'community', 'the'] c Community 2\n",
-      "['d', 'all', 'of', 'these'] a Factors such as living space and temperature range 0\n",
-      "['d', 'all', 'of', 'these'] d All of these 4\n",
-      "['d', 'all', 'of', 'these'] b Where an organism lives 0\n",
-      "['d', 'all', 'of', 'these'] d All of these 4\n",
-      "['a', 'true', 'explanation'] b False 0\n",
-      "['a', 'niche', 'explanation'] a niche 2\n",
-      "['b', 'travelled', 'south'] b Travelled south 3\n",
-      "['c', 'midlatitude'] c Midlatitude cyclone 2\n",
-      "['b', 'false', 'explanation'] a True 0\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['cwww'] c Great Lakes 0\n",
-      "['b', 'false', 'explanation'] a True 0\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['d', 'all', 'of', 'the'] d All of the above 4\n",
-      "['b', 'small', 'pressure', 'difference'] a High pressure gradient between the storm and the air west of the storm 1\n",
-      "['dwww'] d All of the above 0\n",
-      "['c', 'geology'] c Geology 2\n",
-      "['d', 'oceanography'] d Oceanography 2\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['b', 'environmental', 'science'] b Environmental science 3\n",
-      "['a', 'climatology'] d Meteorology Climatology 1\n",
-      "['b', 'oceanography'] a Geological Oceanography 1\n",
-      "['d', 'all', 'of', 'the'] d All of the above 4\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['bwww'] b Barometer 0\n",
-      "['d', 'all', 'of', 'the'] d All of the above 4\n",
-      "['a', 'true', 'explanation'] b False 0\n",
-      "['c', 'meteorologists'] c Meteorologists 2\n",
-      "['d', 'all', 'of', 'these'] a Bouncing radio waves off of the nearest object 1\n",
-      "['dwww'] d All of these 0\n",
-      "['a', 'true'] a True 2\n",
-      "['dwww'] d All of these 0\n",
-      "['b', 'false', 'explanation'] b False 2\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['b', '2', 'relevant'] b Must be tested and have no significant evidence against it 1\n",
-      "['a', 'law', 'explanation'] a Law 2\n",
-      "['a', 'true', 'explanation'] b False 0\n",
-      "['a', 'true', 'explanation'] b False 0\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['b', 'theory', 'explanation'] b Theory 2\n",
-      "['b', 'explanation', 'a'] a Be used to predict the future 1\n",
-      "['b', 'holds', 'under', 'all'] a Does not necessarily hold under all circumstances 2\n",
-      "['b', 'false', 'explanation'] b False 2\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['d', 'greater', 'differences', 'in'] d Greater differences in both daynight and summerwinter temperatures 4\n",
-      "['a', 'the', 'gulf'] a The Gulf Stream 3\n",
-      "['a', 'true', 'explanation'] b False 0\n",
-      "['b', 'false', 'explanation'] b False 2\n",
-      "['a', 'that', 'latitude'] c That latitude has the westerly winds so SF gets weather off the Pacific but VB gets 2\n",
-      "['a', 'the', 'california', 'current'] d A  B 1\n",
-      "['a', 'kansas', 'city'] a Kansas City has a continental climate 3\n",
-      "['b', 'false', 'explanation'] a True 0\n",
-      "['a', 'true', 'explanation'] b False 0\n",
-      "['b', 'false', 'explanation'] a True 0\n",
-      "['a', 'no', 'effect', 'at'] b Dissolved leaves 0\n",
-      "['d', 'explanation', 'l'] d Lichen 1\n",
-      "['cwww'] a Increase precipitation 0\n",
-      "['d', 'h2o'] c NO2 0\n",
-      "['a', 'true'] a True 2\n",
-      "['c', 'causing', 'a', 'die'] c Causing a die off of one species that is then replaced by another 4\n",
-      "['d', 'all', 'of', 'these'] b Mask the amount of global warming that is taking place 1\n",
-      "['d', 'all', 'of', 'the'] d All of the above 4\n",
-      "['b', 'plate', 'tect'] b Plate tectonics 2\n",
-      "['c', 'climate', 'change'] c Climate change 3\n",
-      "['b', 'natural', 'selection'] b Natural selection 3\n",
-      "['a', 'evolution', 'explanation'] a Evolution 2\n",
-      "['b', 'false', 'explanation'] a True 0\n",
-      "['b', 'false', 'explanation'] b False 2\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['b', 'land', 'bridges'] b Land bridges 3\n",
-      "['a', 'evolution', 'the'] a Evolution 2\n",
-      "['c', 'dinosaurs', 'and', 'other'] c Dinosaurs and other organisms at the end of the Mesozoic 4\n",
-      "['c', 'at', 'least'] b A large number of species go extinct in a short amount of time 0\n",
-      "['b', 'false', 'explanation'] a True 0\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['a', 'birds', 'explanation'] a Birds 2\n",
-      "['c', 'species', 'may', 'go'] c Species may go extinct at a slow rate and sometimes all at once 4\n",
-      "['b', 'adaptive', 'radiation'] b Adaptive radiation 3\n",
-      "['d', 'explanation', 'the'] d Any of the above 2\n",
-      "['d', 'all', 'of', 'the'] d All of the above 4\n",
-      "['a', 'food', 'web'] c Food Chain 1\n",
-      "['a', 'true', 'explanation'] b False 0\n",
-      "['a', 'consumers', 'explanation'] d All of the above 0\n",
-      "['dwww'] d All of the above 0\n",
-      "['b', 'false', 'explanation'] b False 2\n",
-      "['a', 'there', 'isnt', 'enough'] a There isnt enough energy to pass along to another trophic level 4\n",
-      "['d', 'all', 'of', 'the'] b Many organisms eat at multiple trophic levels 0\n",
-      "['b', 'false', 'explanation'] a True 0\n",
-      "['c', 'the', 'base', 'of'] c The base of the atmospheric circulation cells 4\n",
-      "['d', 'all', 'of', 'these'] b Causes evaporation 0\n",
-      "['a', 'true', 'explanation'] b False 0\n",
-      "['a', 'the', 'meeting', 'of'] a The meeting of warm moist air from the south and cold dry air from the north 4\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['d', 'all', 'of', 'these'] d All of these 4\n",
-      "['b', 'three', 'circulation', 'cells'] b Three circulation cells in the Northern Hemisphere and three in the Southern 4\n",
-      "['b', 'it', 'is', 'def'] c It is deflected right Northern Hemisphere or left Southern Hemisphere by Coriolis 2\n",
-      "['b', 'false', 'the'] b False 2\n",
-      "['d', 'body', 'temperature'] d Body temperature 3\n",
-      "['a', 'true'] a True 2\n",
-      "['d', 'all', 'of', 'the'] c Swim fly run and walk for transportation 0\n",
-      "['c', 'more', 'fur'] a A high surface area to volume ratio 0\n",
-      "['d', 'wooly', 'mam'] b Whale 0\n",
-      "['d', 'all', 'of', 'the'] d All of the above 4\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['b', 'human', 'activities'] b Human activities such as hunting 3\n",
-      "['b', 'much', 'warmer', 'than'] b Much warmer than today 4\n",
-      "['b', 'false', 'the'] a True 0\n",
-      "['c', 'plants', 'and', 'animals'] c Plants and animals diversified and spread 4\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['b', 'false', 'the'] b False 2\n",
-      "['b', 'evidence', 'that', 'they'] d All of these 0\n",
-      "['b', 'the', 'mesozo'] d The Triassic Epoch 1\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['dwww'] d All of these 0\n",
-      "['a', 'eggs', 'to', 'survive'] a Eggs to survive and hatch away from water 4\n",
-      "['d', 'all', 'of', 'the'] d All of the above 4\n",
-      "['a', 'ordovician'] a Ordovician 2\n",
-      "['c', 'the', 'extent', 'of'] c The extent of shallow seas opened up many marine habitats 4\n",
-      "['b', 'carboniferous'] b Carboniferous seedbearing 2\n",
-      "['a', 'seeds', 'explanation'] a Seeds 2\n",
-      "['b', 'false', 'the'] b False 2\n",
-      "['b', 'changes', 'in', 'the'] c Global warming 0\n",
-      "['b', 'paleozoic'] d Cambrian Explosion Paleozoic 1\n",
-      "['dwww'] b Species composed of calcium carbonate went extinct at a greater rate than others 0\n",
-      "['b', 'false', 'explanation'] b False 2\n",
-      "['b', 'explanation', 'hurricanes'] b Tropical cyclones 1\n",
-      "['d', 'all', 'of', 'the'] d All of the above 4\n",
-      "['b', '3'] c 510 days 0\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['a', 'it', 'goes', 'from'] c It goes from 1 to 5 with 1 being the weakest 3\n",
-      "['a', 'the', 'low', 'pressure'] a The low pressure center of the storm 4\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['d', 'all', 'of', 'these'] d All of these 4\n",
-      "['b', 'it', 'caused', 'the'] b It caused the levees that protected the city to break 4\n",
-      "['a', 'from', 'low', 'pressure'] b From high pressure cells to low pressure cells 3\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['c', 'warm', 'air', 'above'] d Hot air above land rises and sucks in warm wet air from the sea 2\n",
-      "['b', 'hot', 'dry', 'air'] c The Great Basin experiences a high pressure and forces winds back down toward the 0\n",
-      "['d', 'none', 'of', 'these'] a Air descending from the mountains experiences adiabatic heating as it warms it sucks 0\n",
-      "['b', 'valley', 'breezes'] b Valley breezes occur during the day and mountain breezes occur at night 3\n",
-      "['a', 'bring', 'much', 'needed'] a Bring much needed rain for drinking and irrigation 4\n",
-      "['b', 'a', 'high', 'latent'] c A high specific heat compared to land 2\n",
-      "['c', 'occurs', 'in', 'the'] d All of these 0\n",
-      "['c', 'the', 'polar', 'front'] c The polar front large 4\n",
-      "['b', 'fierce', 'n'] b Fierce Noreasters 2\n",
-      "['b', 'air', 'masses', 'blowing'] b Air masses blowing past each other in opposite directions are deflected by Coriolis and 4\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['a', 'the', 'air', 'rises'] a The air rises and cools creating clouds and precipitation 4\n",
-      "['b', 'false', 'explanation'] a True 0\n",
-      "['b', 'false', 'explanation'] b False 2\n",
-      "['c', 'explanation', 'the'] c Storm of the Century 2\n",
-      "['a', 'the', 'tremendous', 'variety'] a The tremendous variety of habitats that organisms have evolved to fill 4\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['b', 'false', 'explanation'] b False 2\n",
-      "['b', 'run', 'away'] b Run away 3\n",
-      "['d', 'all', 'of', 'the'] a About 1 million species are currently alive on Earth 0\n",
-      "['d', 'all', 'of', 'these'] d All of these 4\n",
-      "['c', 'an', 'adaptation'] c An adaptation 3\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['b', 'false', 'explanation'] b False 2\n",
-      "['c', 'gather', 'data', 'using'] c Gather data using scientific method 4\n",
-      "['a', 'satellites', 'explanation'] a Satellites 2\n",
-      "['a', 'independent', 'variable'] a Independent variable 3\n",
-      "['b', 'dependent', 'variable'] b Dependent variable 3\n",
-      "['c', 'the', 'independent', 'factor'] c The independent factor is the time of day and the dependent factor is temperature 4\n",
-      "['b', 'dependent', 'variable'] c Control group 0\n",
-      "['d', 'not', 'be', 'given'] d Not be given vitamin C 4\n",
-      "['b', 'random', 'error'] b Random error 3\n",
-      "['a', 'systematic', 'error'] a Systematic error 3\n",
-      "['b', 'random', 'error'] c Experimental error 1\n",
-      "['d', 'weather', 'model'] b Mathematical model 1\n",
-      "['b', 'meteorologists'] b Meteorologists 2\n",
-      "['b', 'false', 'explanation'] b False 2\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['c', 'we', 'now', 'have'] c We now have much more and much better data 4\n",
-      "['a', 'calculates', 'what', 'will'] d All of these 0\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['dwww'] d All of these 0\n",
-      "['d', 'all', 'of', 'these'] a Allows people to secure their property and evacuate as needed 0\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['a', 'true', 'explanation'] b False 0\n",
-      "['d', 'ear', 'explanation'] d Ear 2\n",
-      "['d', 'explanation', 'the'] d A  B 1\n",
-      "['b', 'the', 'higher', 'pressure'] b The higher pressure on the inside than on the outside would make it push its cap off 4\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['d', 'lack', 'of', 'vegetation'] c Sparser atmosphere 0\n",
-      "['a', 'the', 'gravitational', 'pull'] a The gravitational pull is stronger closer to the center of the Earth 4\n",
-      "['b', 'false', 'explanation'] a True 0\n",
-      "['a', 'true', 'explanation'] b False 0\n",
-      "['d', 'all', 'of', 'these'] d All of these 4\n",
-      "['a', 'producers', 'explanation'] a Producers 2\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['b', 'hyena'] b Hyena 2\n",
-      "['b', 'the', 'pollinator'] d The pollinator gets food and the pollen is spread to other locations 2\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['symbiosis', 'is', 'a', 'relationship'] b A relationship between two interacting species where one or both benefits 2\n",
-      "['c', 'parasitism'] c Parasitism 2\n",
-      "['b', 'false', 'explanation'] a True 0\n",
-      "['d', 'all', 'of', 'the'] d All of the above 4\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['a', 'recommend', 'or', 'deny'] a Recommend or deny a paper for publication 4\n",
-      "['a', 'the', 'community', 'of'] a The community of scientists in the field 4\n",
-      "['b', 'are', 'almost', 'certain'] b Are almost certain to have been done using scientific method 4\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['c', 'data', 'explanation'] c Data 2\n",
-      "['a', 'fact', 'explanation'] a Fact 2\n",
-      "['d', 'bill', 'gates', 'is'] d Bill Gates is the wealthiest man in the United States 4\n",
-      "['a', 'susan', 'has'] a Susan has itchy eyes 3\n",
-      "['d', 'all', 'of', 'the'] d All of the above 4\n",
-      "['a', 'true', 'explanation'] b False 0\n",
-      "['b', 'false', 'explanation'] b False 2\n",
-      "['b', 'each', 'of', 'us'] b Each of us has a guardian angel 4\n",
-      "['b', 'testable'] b Testable 2\n",
-      "['b', 'false', 'the'] b False 2\n",
-      "['a', 'what', 'happens', 'when'] d Do humans and chimpanzees share a common ancestor 1\n",
-      "['a', 'a', 'hypothesis'] d a question 1\n",
-      "['b', 'what', 'is', 'the'] b What is the age of our planet 4\n",
-      "['d', 'all', 'of', 'the'] d All of the above 4\n",
-      "['c', 'i', 'won', 'the'] c I won the lottery because I visualized it happening 4\n",
-      "['b', 'false', 'explanation'] b False 2\n",
-      "['d', 'all', 'of', 'the'] d All of the above 4\n",
-      "['d', 'all', 'of', 'the'] d All of the above 4\n",
-      "['c', 'troposphere'] c Troposphere Stratosphere Mesosphere Thermosphere 2\n",
-      "['a', 'temperature', 'gradient'] a Temperature gradient 3\n",
-      "['b', 'stratosphere'] a Troposphere 0\n",
-      "['a', 'cool', 'gas', 'molecules'] d Warm gas molecules move more and take up more space so they become less dense 2\n",
-      "['a', 'troposphere'] a Troposphere 2\n",
-      "['a', 'true'] a True 2\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['b', 'the', 'heat', 'source'] b The heat source of that layer 4\n",
-      "['a', 'from', 'severe', 'thunder'] a From severe thunderstorms 3\n",
-      "['c', 'wind', 'speed', 'and'] c Wind speed and damage 4\n",
-      "['d', 'cumulon'] d cumulonimbus 1\n",
-      "['b', 'very', 'few', 'tornado'] c Tornadoes strike a very wide path up to several miles 1\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['b', 'false', 'explanation'] b False 2\n",
-      "['a', 'maritime', 'tropical', 'and'] b Maritime tropical and continental polar 3\n",
-      "['b', 'false', 'explanation'] b False 2\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['c', 'high', 'because', 'there'] c High because there is a lot of difference in temperature salinity light and other factors 4\n",
-      "['b', 'plankton'] b Plankton 2\n",
-      "['a', 'true', 'explanation'] b False 0\n",
-      "['a', 'true', 'explanation'] b False 0\n",
-      "['b', 'false', 'explanation'] a True 0\n",
-      "['b', 'false', 'explanation'] b False 2\n",
-      "['a', 'in', 'the', 'phot'] b Close to shore 0\n",
-      "['c', 'account', 'for', 'a'] c Account for a lot of the oceans biodiversity 4\n",
-      "['d', 'all', 'of', 'these'] d All of these 4\n",
-      "['d', 'all', 'of', 'these'] a They have fur fat fast metabolisms small surface area to volume for warmth 0\n",
-      "['a', 'meet', 'a'] a Meet 2\n",
-      "['b', 'have', 'different', 'densities'] b Have different densities so the less dense one goes up over the denser one 4\n",
-      "['awww'] c Break apart after several days 0\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['a', 'explanation', 'a'] d A squall line 1\n",
-      "['d', 'explanation', 'a'] a Thunderstorms 1\n",
-      "['b', 'a', 'warm', 'air'] b A warm air mass slides over a cold air mass 4\n",
-      "['b', 'warm', 'air', 'and'] d Cold air and snow give way to warmer sleet and freezing rain with stratus clouds and 2\n",
-      "['a', 'true'] a True 2\n",
-      "['a', 'true'] b False 0\n",
-      "['b', 'amount', 'of', 'precipitation'] d Barometric pressure at the elevation of each location 1\n",
-      "['a', 'sea', 'level'] a Sea level 3\n",
-      "['b', 'explanation', 'jet'] b Jet streams 2\n",
-      "['a', 'hurricane'] a Hurricane 2\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['cwww'] c A front 0\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['b', 'false', 'explanation'] b False 2\n",
-      "['dwww'] c A low pressure cell 0\n",
-      "['d', 'all', 'of', 'these'] d All of these 4\n",
-      "['b', 'false', 'explanation'] b False 2\n",
-      "['b', 'false', 'explanation'] b False 2\n",
-      "['d', 'all', 'of', 'these'] d All of these 4\n",
-      "['a', 'humidity'] a Humidity 2\n",
-      "['b', 'false', 'explanation'] a True 0\n",
-      "['b', 'false', 'explanation'] b False 2\n",
-      "['c', 'changeable'] c Changeable 2\n",
-      "['b', 'the', 'amount', 'of'] b The amount of solar radiation it gets 4\n",
-      "['b', 'convection'] b Convection 2\n",
-      "['b', 'outward', 'from', 'a'] a Toward a zone of rising air 1\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['b', 'false', 'explanation'] b False 2\n",
-      "['d', 'all', 'of', 'the'] d All of the above 4\n",
-      "['d', 'all', 'of', 'these'] c Windmills are expensive to build and maintain 0\n",
-      "['c', 'warm', 'air', 'in'] b Warm air from the ocean rises and sucks cool air across the pass from the ocean 2\n",
-      "['d'] a Unattractive 0\n",
-      "['b', 'false', 'the'] b False 2\n",
-      "['a', 'true', 'explanation'] a True 2\n",
-      "['b', 'theory', 'explanation'] b theory 2\n",
-      "['dwww'] d all of the above 0\n",
-      "['b', 'tested', 'and', 'confirmed'] b tested and confirmed repeatedly 4\n",
-      "['dwww'] d all of the above 0\n",
-      "['bwww'] b repeatedly tested and supported by the results 0\n",
-      "['d', 'explanation', 'the'] d making observations of wildlife while hiking in the woods 2\n",
-      "['d', 'all', 'of', 'the'] d all of the above 4\n",
-      "['dwww'] d evolving 0\n",
-      "['a', 'think', 'of', 'a'] a think of a way to test the idea 4\n",
-      "['a', 'wondering', 'why', 'fire'] a wondering why fire flies produce light 4\n",
-      "['cwww'] c to make sure the results are reliable 0\n",
-      "['c', 'put', 'the', 'idea'] c put the idea aside until it can be tested or replace it with an idea that can be tested 4\n",
-      "['d', 'science'] dscience 0\n",
-      "['f', 'explanation', 'a'] greasoning 0\n",
-      "['fwww'] fscientific theory 0\n",
-      "['cwww'] cscientist 0\n",
-      "['b', 'false', '__'] b false 2\n",
-      "['a', 'true', '__'] b false 0\n",
-      "['b', 'evidence', 'answer'] bevidence 0\n",
-      "['a', 'true', '__'] a true 2\n",
-      "['a', 'scientific', 'law'] ascientific law 1\n",
-      "['ewww'] eskepticism 0\n",
-      "['a', 'true', '__'] b false 0\n",
-      "['b', 'false', '__'] b false 2\n",
-      "['a', 'true', '________'] b false 0\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['b', 'false', 'science'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['d', 'more', 'than'] d more than 1000000 3\n",
-      "['a', 'at', 'least', 'one'] a at least one cell 4\n",
-      "['cwww'] c paleontologist 0\n",
-      "['c', 'bacteria', 'the'] c bacteria 2\n",
-      "['b', 'theory', 'of', 'evolution'] d two of the above 1\n",
-      "['c', 'studying', 'rain', 'forest'] a studying yeast cells to learn how they divide 1\n",
-      "['a', 'finding', 'solutions', 'to'] a finding solutions to practical problems 4\n",
-      "['f', 'cell'] fcell 0\n",
-      "['a', 'basic', 'science'] capplied science 1\n",
-      "['e', 'life', 'science'] gorganism 0\n",
-      "['a', 'basic', 'science'] abasic science 1\n",
-      "['b', 'ecology'] becology 0\n",
-      "['e', 'life', 'science'] elife science 1\n",
-      "['dwww'] devolution 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['bwww'] b method 0\n",
-      "['c', 'scientific', 'method'] c scientific method 3\n",
-      "['d', 'all', 'of', 'the'] d all of the above 4\n",
-      "['c', 'tested'] d testable 0\n",
-      "['b', 'testing', 'a', 'hypothesis'] d two of the above 0\n",
-      "['dwww'] a false when it really is false 0\n",
-      "['d', 'all', 'of', 'the'] d all of the above 4\n",
-      "['awww'] a hypothesis 0\n",
-      "['bwww'] b dependent variable 0\n",
-      "['a', 'observations', 'a'] a observations 2\n",
-      "['b', 'provide', 'support', 'for'] b provide support for the hypothesis 4\n",
-      "['dwww'] d all of the above 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['a', 'true', '________'] a true 2\n",
-      "['b', 'false', 'the'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true'] a true 2\n",
-      "['a', 'true', 'the'] b false 0\n",
-      "['b', 'control'] bcontrol 0\n",
-      "['dwww'] dobservation 0\n",
-      "['ewww'] gindependent variable 0\n",
-      "['a', 'hypothesis'] ahypothesis 0\n",
-      "['f', 'replication'] freplication 0\n",
-      "['ewww'] edependent variable 0\n",
-      "['cwww'] cexperiment 0\n",
-      "['b', 'biohazard'] ehigh heat 0\n",
-      "['b', 'biohazard'] csharp instrument 0\n",
-      "['b', 'biohazard'] aelectrical hazard 0\n",
-      "['b', 'biohazard'] glaser radiation 0\n",
-      "['b', 'biohazard'] dexplosive substance 0\n",
-      "['b', 'biohazard'] bbiohazard 0\n",
-      "['b', 'biohazard'] fchemical hazard 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['a', 'true'] a true 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['dwww'] d all of the above 0\n",
-      "['dwww'] b poisonous chemical 0\n",
-      "['dwww'] d Clean up any spills immediately 0\n",
-      "['b', 'fan', 'vapors'] b Fan vapors toward your nose rather than smelling them directly 3\n",
-      "['c', 'wash', 'it'] d FIGURE 11 0\n",
-      "['a', 'long', 'sleeves'] b a hazmat suit 1\n",
-      "['d', 'all', 'of', 'the'] d all of the above 4\n",
-      "['b', 'false', '________'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['dwww'] d produces rings inside the trunk of the tree 0\n",
-      "['cwww'] c form from haploid spores 0\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['bwww'] b vascular tissue 0\n",
-      "['b', 'false', '________'] b false 2\n",
-      "['dwww'] d all of the above 0\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['b', 'false', 'the'] a true 0\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['a', 'true'] b false 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['f'] fvascular 0\n",
-      "['c'] ccuticle 0\n",
-      "['d'] gmeristem 0\n",
-      "['f', 'vascular'] dground 0\n",
-      "['e', 'stoma'] estoma 0\n",
-      "['a', 'dermal'] adermal 0\n",
-      "['b', 'chloroplast'] bchloroplast 0\n",
-      "['d', 'all', 'of', 'the'] d all of the above 4\n",
-      "['cwww'] c dermal tissue 0\n",
-      "['awww'] c releasing oxygen 0\n",
-      "['b'] d two of the above 0\n",
-      "['dwww'] d all of the above 0\n",
-      "['a', 'are', 'plants', 'in'] a are plants in the diploid generation 4\n",
-      "['bwww'] d none of the above 0\n",
-      "['a', 'stonew'] a stoneworts 1\n",
-      "['c'] cangiosperm 0\n",
-      "['d'] dcone 0\n",
-      "['a', 'vascular', 'tissues'] a vascular tissues 3\n",
-      "['cwww'] c germination 0\n",
-      "['b'] bstamen 0\n",
-      "['d'] egymnosperm 0\n",
-      "['d', 'ovary'] d ovary 2\n",
-      "['a', 'seed'] aseed 0\n",
-      "['bwww'] b mosses 0\n",
-      "['f'] fbryophyte 0\n",
-      "['b'] gpistil 0\n",
-      "['a', 'evolved', 'from', 'green'] a evolved from green algae 4\n",
-      "['cwww'] c has a backup copy of each gene 0\n",
-      "['a', 'vascular', 'plants'] a Vascular plants evolved after seed plants had appeared 3\n",
-      "['dwww'] d all of the above 0\n",
-      "['a', 'true', 'the'] a true 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['bwww'] b pollen tube 0\n",
-      "['b', 'false', 'v'] b false 2\n",
-      "['b', 'attracting', 'pollin'] b attracting pollinators 2\n",
-      "['a', 'ferns'] a ferns 2\n",
-      "['a', 'true'] a true 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true'] a true 2\n",
-      "['b', 'false', 'the'] b false 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['b', 'false', 'the'] b false 2\n",
-      "['a', 'true'] b false 0\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true'] b false 0\n",
-      "['c', 'turning', 'toward', 'or'] c turning toward or away from a stimulus 4\n",
-      "['b', 'false', 'pl'] b false 2\n",
-      "['a', 'true'] a true 2\n",
-      "['c', 'kill', 'bacteria'] c kill bacteria 3\n",
-      "['a', 'have', 'special', 'adaptations'] a have special adaptations 4\n",
-      "['a', 'true'] b false 0\n",
-      "['true'] b false 0\n",
-      "['awww'] a increase water intake 0\n",
-      "['b', 'false', 'c'] b false 2\n",
-      "['dwww'] d all of the above 0\n",
-      "['true'] a true 1\n",
-      "['a', 'true'] b false 0\n",
-      "['cwww'] cxerophyte 0\n",
-      "['dwww'] ddormancy 0\n",
-      "['e'] ggravitropism 0\n",
-      "['a', 'adaptation'] aadaptation 0\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['e', 'tropism'] etropism 0\n",
-      "['a', 'true'] a true 2\n",
-      "['f'] fepiphyte 0\n",
-      "['b'] bphototropism 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['cwww'] c roots 0\n",
-      "['cwww'] c a light source 0\n",
-      "['d', 'nutrients', 'explanation'] d nutrients 2\n",
-      "['dwww'] d stimulus 0\n",
-      "['a', 'they', 'have', 'ancestors'] a They have ancestors that lived on land 4\n",
-      "['dwww'] d all of the above 0\n",
-      "['bwww'] b support 0\n",
-      "['dwww'] a specialized cells 0\n",
-      "['d', 'all', 'of', 'the'] d all of the above 4\n",
-      "['bwww'] b insect larva 0\n",
-      "['a', 'earthworms'] d snakes 0\n",
-      "['bwww'] b a cell wall 0\n",
-      "['a', 'true'] a true 2\n",
-      "['a', 'true', 'the'] a true 2\n",
-      "['b', 'false', '_an'] b false 2\n",
-      "['a', 'true'] b false 0\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['b', 'false', '_'] a true 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true', 'the'] b false 0\n",
-      "['b', 'chordata'] bChordata 0\n",
-      "['b', 'chordata'] glarva 0\n",
-      "['b', 'chordata'] dvertebral column 0\n",
-      "['b', 'chordata'] cEukarya 0\n",
-      "['a', 'vertebrate'] avertebrate 0\n",
-      "['ewww'] ezygote 0\n",
-      "['f', 'fertilization'] ffertilization 0\n",
-      "['a', 'sponges'] a sponges 2\n",
-      "['dwww'] d all of the above 0\n",
-      "['c', 'respond', 'to', 'stimuli'] a move on their own 0\n",
-      "['cwww'] d two of the above 0\n",
-      "['a', 'chloroplast'] a chloroplast 2\n",
-      "['awww'] a send signals to other cells 0\n",
-      "['dwww'] d all of the above 0\n",
-      "['b', 'between', '450'] b between 450 and 400 million years ago 3\n",
-      "['a', 'true', 'the'] a true 2\n",
-      "['a', 'true'] b false 0\n",
-      "['awww'] a cell specialization  body symmetry body segmentation 0\n",
-      "['dwww'] d all of the above 0\n",
-      "['b', 'false'] a true 0\n",
-      "['b', 'false', 'a'] b false 2\n",
-      "['cwww'] a sponge 0\n",
-      "['a', 'true'] a true 2\n",
-      "['d', 'two', 'of', 'the'] d two of the above 4\n",
-      "['b', 'false'] a true 0\n",
-      "['a', 'true'] b false 0\n",
-      "['b', 'false', 'the'] a true 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true'] a true 2\n",
-      "['d', 'segmentation'] dsegmentation 0\n",
-      "['e', 'chordate'] echordate 0\n",
-      "['f', 'mesoderm'] fmesoderm 0\n",
-      "['c', 'coelom'] ccoelom 0\n",
-      "['e', 'chordate'] gamniote 0\n",
-      "['b', 'symmetry'] bsymmetry 0\n",
-      "['a', 'notochord'] anotochord 0\n",
-      "['b', 'about', '550'] a at least 630 million years ago 0\n",
-      "['b', 'coelom'] a body symmetry 0\n",
-      "['bwww'] b sponges 0\n",
-      "['dwww'] a beetle 0\n",
-      "['dwww'] d all of the above 0\n",
-      "['dwww'] d roundworms 0\n",
-      "['a', 'coelom'] a coelom 2\n",
-      "['cwww'] c legs 0\n",
-      "['dwww'] d all of the above 0\n",
-      "['bwww'] b a flexible endoskeleton 0\n",
-      "['cwww'] c there is no distinct larva stage 0\n",
-      "['cwww'] c internal organs are located 0\n",
-      "['cwww'] c snails 0\n",
-      "['cwww'] c multiple body segments 0\n",
-      "['a', 'shed', 'an', 'out'] a shed an outgrown exoskeleton 3\n",
-      "['dwww'] d all of the above 0\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['c', 'explanation', 'in'] c Most insect species have already been identified 1\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['dwww'] d all of the above 0\n",
-      "['dwww'] d fly 0\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['b', 'false', 'the'] b false 2\n",
-      "['a', 'true'] a true 2\n",
-      "['a', 'true'] a true 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['a', 'true'] a true 2\n",
-      "['f'] fmolting 0\n",
-      "['f'] gpupa 0\n",
-      "['d'] aInsecta 0\n",
-      "['b'] bmetamorphosis 0\n",
-      "['ewww'] ethorax 0\n",
-      "['d'] dArthropoda 0\n",
-      "['c'] ccuticle 0\n",
-      "['bwww'] b tunicates 0\n",
-      "['f'] bbilateral 0\n",
-      "['c', 'chordata'] cChordata 0\n",
-      "['bwww'] b radial symmetry as adults 0\n",
-      "['awww'] a a central nervous system 0\n",
-      "['c', 'chordata'] eechinoderm 0\n",
-      "['f'] fradial 0\n",
-      "['cwww'] d two of the above 0\n",
-      "['e'] gEchinodermata 0\n",
-      "['dwww'] b the human species 0\n",
-      "['d'] dtube foot 0\n",
-      "['a', 'notochord'] anotochord 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['b', 'false', 'the'] b false 2\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['b', 'false', 'there'] a true 0\n",
-      "['a', 'true'] a true 2\n",
-      "['b', 'false', 'the'] b false 2\n",
-      "['a', 'true'] b false 0\n",
-      "['a', 'true'] a true 2\n",
-      "['a', 'true'] a true 2\n",
-      "['a', 'true', 'the'] a true 2\n",
-      "['a', 'true'] a true 2\n",
-      "['cwww'] d two of the above 0\n",
-      "['dwww'] d all of the above 0\n",
-      "['awww'] a heart for pumping blood 0\n",
-      "['dwww'] c fissioning 0\n",
-      "['dwww'] d all of the above 0\n",
-      "['a', 'more', 'than'] a more than 500 million years ago 3\n",
-      "['cwww'] c They lose their notochord as adults 0\n",
-      "['b', 'four', 'legs'] b four legs 3\n",
-      "['d', 'all', 'of', 'the'] d all of the above 4\n",
-      "['dwww'] d frogs 0\n",
-      "['dwww'] a algae 0\n",
-      "['c', 'amphibians'] d two of the above 0\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true'] a true 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['a', 'true'] a true 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true'] a true 2\n",
-      "['b', 'false', 'am'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['b', 'false', 'ca'] a true 0\n",
-      "['b', 'false', 'am'] b false 2\n",
-      "['c'] esalamander 0\n",
-      "['b'] bmucus 0\n",
-      "['a', 'lungs'] alungs 0\n",
-      "['f'] fcaecilian 0\n",
-      "['b'] gkeratin 0\n",
-      "['c'] ccloaca 0\n",
-      "['d'] dgills 0\n",
-      "['cwww'] d scales 0\n",
-      "['b', 'fertilization'] b Fertilization may take place inside or outside the body 2\n",
-      "['dwww'] d all of the above 0\n",
-      "['cwww'] a develops lungs 0\n",
-      "['bwww'] b Antarctica 0\n",
-      "['dwww'] d all of the above 0\n",
-      "['awww'] a saltwater lakes 0\n",
-      "['c', 'reduce', 'water', 'loss'] c reduce water loss 4\n",
-      "['f', 'squamata'] fSquamata 0\n",
-      "['e'] eomnivore 0\n",
-      "['b', 'modern', 'reptiles'] d The Testudines Order includes turtles and lizards 0\n",
-      "['d', 'all', 'of', 'the'] d all of the above 4\n",
-      "['g', 'testudines'] gTestudines 0\n",
-      "['d'] dcarnivore 0\n",
-      "['dwww'] c lizards 0\n",
-      "['a', 'sphenodont'] aSphenodontia 0\n",
-      "['d'] b turtle  leaves 0\n",
-      "['c'] cCrocodilia 0\n",
-      "['b'] bdiaphragm 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['b', 'false', 'the'] b false 2\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['b', 'false', 'sn'] a true 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['cwww'] c salamanders 0\n",
-      "['dwww'] d all of the above 0\n",
-      "['b', 'explanation', 're'] b Reptiles release sperm or eggs into fresh water 1\n",
-      "['cwww'] c Squamata Order 0\n",
-      "['bwww'] a alligators and crocodiles 0\n",
-      "['dwww'] d Testudines Order 0\n",
-      "['c', 'carnivores'] c carnivores 2\n",
-      "['bwww'] b flying 0\n",
-      "['dwww'] b are bipedal 0\n",
-      "['b', 'explanation', 'birds'] d Like reptiles birds lay amniotic eggs and are ectothermic 1\n",
-      "['c', 'to', 'attract', 'mates'] c to attract mates 4\n",
-      "['c', 'killdeers'] a robins 0\n",
-      "['d', 'all', 'of', 'the'] d all of the above 4\n",
-      "['a', 'reproduction', 'can'] b Separate sexes produce sperm and eggs 0\n",
-      "['dwww'] d honeyeaters 0\n",
-      "['b', '7'] d 29 0\n",
-      "['a', 'specialist', 'explanation'] b omnivore 0\n",
-      "['dwww'] a moa 0\n",
-      "['d', 'all', 'of', 'the'] d all of the above 4\n",
-      "['b', 'false', 'the'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['b', 'false', 'bird'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['b', 'false', 'the'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['f'] fgeneralist 0\n",
-      "['e'] eincubation 0\n",
-      "['d'] dspecialist 0\n",
-      "['f', 'generalist'] gperching bird 0\n",
-      "['b'] bbipedal 0\n",
-      "['a'] anocturnal raptor 0\n",
-      "['c'] ccourtship 0\n",
-      "['cwww'] d two of the above 0\n",
-      "['cwww'] c echidnas 0\n",
-      "['bwww'] b increasing their metabolic rate 0\n",
-      "['c', 'manatee'] c manatee 2\n",
-      "['c', 'rodentia'] c Rodentia 2\n",
-      "['a', 'true', 'the'] a true 2\n",
-      "['a', 'true', 'the'] a true 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['b', 'false', 'there'] a true 0\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['b', 'false', 'the'] b false 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['b', 'false', 'the'] b false 2\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['dwww'] d all of the above 0\n",
-      "['cwww'] c cheetah 0\n",
-      "['a', 'include', 'chimpanze'] a include chimpanzees 2\n",
-      "['a', 'monotreme', 'mammals'] a monotreme mammals 3\n",
-      "['dwww'] b echidnas 0\n",
-      "['cwww'] c monotreme mammals 0\n",
-      "['dwww'] b marsupials 0\n",
-      "['e'] emarsupial mammal 0\n",
-      "['f'] fsweat 0\n",
-      "['e'] gmonotreme mammal 0\n",
-      "['b'] bplacenta 0\n",
-      "['d'] dlactation 0\n",
-      "['b', 'placenta'] cplacental mammal 0\n",
-      "['a'] aalveolus 0\n",
-      "['c'] c flying squirrel 1\n",
-      "['b', 'false'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['c', 'gorilla'] c gorilla 2\n",
-      "['b', 'vision'] b vision 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['b', 'high', 'levels', 'of'] b high levels of intelligence 4\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['c', 'fruit'] c fruit 2\n",
-      "['b', 'false', 'ch'] b false 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['b', 'false', 'pr'] b false 2\n",
-      "['b', 'false', 'pro'] b false 2\n",
-      "['e'] estereoscopic 0\n",
-      "['d'] darboreal 0\n",
-      "['f'] avervet 0\n",
-      "['f'] fprosimians 0\n",
-      "['b'] bbrachiation 0\n",
-      "['c'] csquirrel monkey 0\n",
-      "['d'] gopposable 0\n",
-      "['bwww'] b koalas 0\n",
-      "['cwww'] c trees 0\n",
-      "['a', 'larger', 'brains'] a larger brains 3\n",
-      "['bwww'] c omnivores 0\n",
-      "['c', 'explanation', 'pro'] c They include New World and Old World monkeys 1\n",
-      "['bwww'] b other fingers on the same hand 0\n",
-      "['dwww'] a tropical rainforests 0\n",
-      "['dwww'] d learning to avoid eating certain foods because they taste bad 0\n",
-      "['dwww'] d all of the above 0\n",
-      "['bwww'] b predictable 0\n",
-      "['dwww'] d insight learning 0\n",
-      "['bwww'] b more flexible 0\n",
-      "['awww'] d using a twig as a tool 0\n",
-      "['cwww'] c reflex behavior 0\n",
-      "['cwww'] d reflex learning 0\n",
-      "['dwww'] d all of the above 0\n",
-      "['bwww'] b innate behavior 0\n",
-      "['b', 'conditioning'] b conditioning 2\n",
-      "['d', 'both', 'mammals', 'and'] d both mammals and birds 4\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['b', 'false', 'the'] a true 0\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true'] b false 0\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['e', 'conditioning'] flearned behavior 0\n",
-      "['e', 'conditioning'] econditioning 0\n",
-      "['d'] danimal behavior 0\n",
-      "['e', 'conditioning'] ginsight learning 0\n",
-      "['e', 'conditioning'] creflex behavior 0\n",
-      "['e', 'conditioning'] ahabituation 0\n",
-      "['b'] binnate behavior 0\n",
-      "['d', 'all', 'of', 'the'] d all of the above 4\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['bwww'] b ants will be able to find it 0\n",
-      "['b', 'mate', 'with', 'me'] a Stay away 0\n",
-      "['b', 'false', 'the'] b false 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['d', 'all', 'of', 'the'] d all of the above 4\n",
-      "['b', 'false', 'anim'] a true 0\n",
-      "['c', 'workers'] c workers 2\n",
-      "['b', 'false', 'explanation'] a true 0\n",
-      "['b', 'false', 'the'] b false 2\n",
-      "['b'] flanguage 0\n",
-      "['b'] gdiurnal 0\n",
-      "['e'] esocial animal 0\n",
-      "['b'] dnocturnal 0\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['c'] ccommunication 0\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['b'] bcircadian rhythm 0\n",
-      "['a'] abiological clock 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['dwww'] d any of the above 0\n",
-      "['bwww'] b They mark the trail to the food source with chemicals 0\n",
-      "['c', 'language', 'the'] c language 2\n",
-      "['dwww'] d all of the above 0\n",
-      "['a', 'releasing', 'chemicals', 'in'] a releasing chemicals in urine 4\n",
-      "['bwww'] b All birds migrate south for the winter 0\n",
-      "['d', 'day'] d day 2\n",
-      "['b', 'fruits'] b fruits 2\n",
-      "['b'] bMyPlate 0\n",
-      "['a'] gmain ingredient 0\n",
-      "['cwww'] a Fill half of your plate fruits and vegetables 0\n",
-      "['dwww'] a serving size 0\n",
-      "['d'] dnutrition facts label 0\n",
-      "['e'] eobesity 0\n",
-      "['b', '10', 'percent'] d 20 percent 1\n",
-      "['f'] fserving size 0\n",
-      "['d', 'all', 'of', 'the'] d all of the above 4\n",
-      "['c'] cpercent daily value 0\n",
-      "['a'] aingredient 0\n",
-      "['a', 'true', 'explanation'] b false 0\n",
-      "['awww'] a true 0\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['bwww'] a true 0\n",
-      "['a', 'true'] a true 2\n",
-      "['b', 'false', 'explanation'] b false 2\n",
-      "['b', 'false', 'the'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['b', 'false', 'the'] b false 2\n",
-      "['a', 'true', 'explanation'] a true 2\n",
-      "['b', 'grains', 'and', 'dairy'] a fruits and vegetables 1\n",
-      "['d', 'all', 'of', 'the'] c pasta 0\n",
-      "['dwww'] d all of the above 0\n",
-      "['c', 'whole', 'grains'] c whole grains 3\n",
-      "['d', 'body', 'mass', 'index'] d body mass index 4\n",
-      "['b', 'physical', 'activity', 'is'] d You should get about 15 minutes of physical activity per day 2\n",
-      "['dwww'] d all of the above 0\n",
-      "['2', 'colá'] 2 Colágeno 1\n",
-      "['1', 'mito'] 1 Mitocondrial interna 1\n",
-      "['4', 'locom'] 4 Locomoción celular 1\n",
-      "['2', 'mito'] 4 Endosomas 0\n",
-      "['2', 'cél'] 2 Células de Schwann 1\n",
-      "['3', 'discos'] 3 Discos intervertebrales 2\n",
-      "['1', 'nucle'] 3 Cromosoma eucariótico 0\n",
-      "['1', 'condro'] 4 Osteoclastos 0\n",
-      "['1', 'conjunt'] 4 Cartílago 0\n",
-      "['1', 'intram'] 3 Endocondral 0\n",
-      "['1', 'proteos'] 1 Proteosomas 1\n",
-      "['2'] 1 Retículo endoplásmico 0\n",
-      "['3', 'arp'] 3 Arp 23 2\n",
-      "['1', 'axon'] 2 Centrosoma 0\n",
-      "['2', 'cromát'] 4 Cromosomas homólogos 0\n",
-      "['3', 'ácido'] 3 Ácido hialurónico 2\n",
-      "['1', 'neutr'] 1 Neutrófilo 1\n",
-      "['2', 'conjunt'] 2 Conjuntivo 1\n",
-      "['2'] 2 Cardiaco 1\n",
-      "['1', 'cere'] 4 Médula espinal 0\n",
-      "['4', 'explanation'] 4 Una de las funciones de las proteínas transmembrana es la recepción de ligandos 1\n",
-      "['1', 'f'] 2 Los cariotipos clásicos se confeccionan con cromosomas metafásicos 0\n",
-      "['1', '2'] 2 Presentan un canal acuoso 1\n",
-      "['1', 'cines'] 3 Los microtúbulos se forman por polimerización 0\n",
-      "['2', 'explanation'] 2 Las vías urinarias es el de transición 1\n",
-      "['1', '2'] 1 Epitelial 1\n",
-      "['1', 'lax'] 3 Denso presenta escasas células y abundantes fibras 0\n",
-      "['1', 'f'] 2 La función más importante del adiposo pardo es la termogénesis 0\n",
-      "['1', 'f'] 4 El cartílago presenta dos modalidades de crecimiento por aposición e intersticial 0\n",
-      "['3', 'pán'] 3 Páncreas exocrino 1\n",
-      "['1', 'cara'] 1 Cara 2\n",
-      "['3', 'muscar'] 4 Nicotínicos 0\n",
-      "['3'] 1 Propagan el potencial de acción al interior celular 0\n",
-      "['1', 'olf'] 1 Olfato 1\n",
-      "['2', 'bicon'] 2 Biconvexa 1\n",
-      "['3'] 3 Participan en la planificación de los movimientos voluntarios 1\n",
-      "['4', 'efer'] 2 Células horizontales 0\n",
-      "['4', 'frontal'] 2 Temporal 0\n",
-      "['4', 'explanation'] 3 Cada fibra recibe inervación de una sola motoneurona 0\n",
-      "['1', '2'] 4 Son receptores de adaptación rápida 0\n",
-      "['2'] 1 Monitoriza cambios en la tensión muscular 0\n",
-      "['3', 'dienc'] 3 Diencéfalo 1\n",
-      "['1', 'insul'] 1 Insulina 1\n",
-      "['1', 'sí'] 4 Síntesis y liberación de corticoliberina CRH 0\n",
-      "['1', 'incorporación'] 2 Neoglucogénesis 0\n",
-      "['2', '23'] 1 Parathormona 0\n",
-      "['3'] 3 Son inmunosupresores 1\n",
-      "['3', 'estero'] 1 Tróficas 0\n",
-      "['1', '2'] 1 Está modulada por el vago 1\n",
-      "['1', '2'] 4 Inhibe la contracción del cuerpo y antro gástricos 0\n",
-      "['2', 'trips'] 2 Tripsinógeno 1\n",
-      "['3'] 3 Reflejo gastrocólico 1\n",
-      "['3', 'explanation'] 1 Determina la diferenciación del sexo masculino 0\n",
-      "['1', '2'] 1 Estimula la secreción de andrógenos por las células de Leydig 1\n",
-      "['1', 'formación'] 3 Fibrinolisis 0\n",
-      "['3', 'volum'] 4  en volumen de glóbulos rojos respecto a sangre 0\n",
-      "['1', 'hay', 'un'] 1 Hay un retraso en el nódulo aurículoventricular 3\n",
-      "['2', 'nodo', 'sin'] 2 Nodo sinusal 2\n",
-      "['1', '2'] 2 Aumento de la concentración de hormona tiroidea 1\n",
-      "['2'] 2 Las válvulas aurículoventriculares están cerradas 1\n",
-      "['1', 'estim'] 2 Secreción de ADH 0\n",
-      "['3', 'el', 'interval'] 3 El intervalo PR 2\n",
-      "['3', 'vagal'] 4 Simpática 0\n",
-      "['1', '2'] 2 El gas espirado proviene del espacio muerto anatómico 1\n",
-      "['2', 'durante'] 2 Durante la inspiración 2\n",
-      "['3', 'la', 'dist'] 3 La distensibilidad pulmonar 2\n",
-      "['2'] 2 Se intercambia con el exterior en una respiración normal 1\n",
-      "['4', 'explanation'] 3 Posee una envoltura lipídica 0\n",
-      "['2', 'cada'] 2 Cada una de las proteínas que constituyen la cápside de un virus 2\n",
-      "['3', 'arn'] 3 ARN de cadena sencilla  2\n",
-      "['1', 'inhib'] 4 Inhibe la proteasa 0\n",
-      "['3', 'acic'] 3 Aciclovir 1\n",
-      "['1', 'flav'] 1 Flavivirus 1\n",
-      "['2', 'tog'] 4 Rabdovirus 0\n",
-      "['2', 'ortom'] 2 Ortomixovirus 1\n",
-      "['2', 'citome'] 2 Citomegalovirus 1\n",
-      "['3', 'hepad'] 3 Hepadnavirus 1\n",
-      "['2', 'virus', 'de'] 1 Papilomavirus 0\n",
-      "['2', 'papil'] 3 Parvovirus 0\n",
-      "['3', 'cuer'] 3 Cuerpos de Negri 1\n",
-      "['3', 'pato'] 4 Cerdo 0\n",
-      "['4', 'desin'] 1 Antiséptico 0\n",
-      "['2', '0'] 2 022 μm 1\n",
-      "['3', 'el', 'ad'] 3 El ADN 2\n",
-      "['1', 'disol'] 2 Alquilante 0\n",
-      "['3'] 4 Son agentes reductores 0\n",
-      "['1', 'lisoz'] 1 Lisozima 1\n",
-      "['2'] 2 Endosporas 1\n",
-      "['1', 'lí'] 1 Lípido A 1\n",
-      "['3', 'fts'] 3 FtsZ 1\n",
-      "['3'] 1 Secuencia de inserción 0\n",
-      "['2', 'mil', 'veces'] 4 Mil veces más que la longitud de la bacteria 2\n",
-      "['3'] 4 Unirse al ADN protegiéndolo del calor y la radiación ultravioleta 0\n",
-      "['4', 'transpe'] 4 Transpeptidación 1\n",
-      "['1', 'sulf'] 1 Sulfanilamida 1\n",
-      "['2', 'las', 'flu'] 2 Las fluoroquinolonas 2\n",
-      "['3'] 3 Se transfieren aquellos que contienen el operón tra y el sitio oriT 1\n",
-      "['1'] 3 Realiza la fermentación butilén glicólica 23 butanodiol 0\n",
-      "['2', 'mycop'] 2 Mycoplasma 1\n",
-      "['4', 'staph'] 4 Staphylococcus 1\n",
-      "['1'] 1 Pulgas 1\n",
-      "['2', 'coxi'] 2 Coxiella burnetti  fiebre Q 1\n",
-      "['2', 'micob'] 3 Rickettsias 0\n",
-      "['2', 'ure'] 2 Ureasa 1\n",
-      "['1', 'strept'] 1 Streptococcus 1\n",
-      "['3'] 4 Microscopía de campo oscuro 0\n",
-      "['2'] 3 Ser ácidoalcohol resistentes 0\n",
-      "['3', 'isla'] 4 Profago atenuado 0\n",
-      "['1', 'citoc'] 1 Citocromo c oxidasa 1\n",
-      "['2', 'la', 'pres'] 1 El mayor contenido en peptidoglicano 0\n",
-      "['2'] 2 Agar sabourauddextrosa 1\n",
-      "['1', 'sí'] 1 Síntesis de ergosterol 1\n",
-      "['1', 'thrich'] 1 Thrichophyton 1\n",
-      "['4', 'asper'] 4 Aspergillus flavus 1\n",
-      "['www'] 3 Puede formar parte de la microbiota normal del cuerpo humano 0\n",
-      "['1'] 1 Incrementa la expresión de moléculas MHC de clase II 1\n",
-      "['4'] 2 Procesan el antígeno a través de la vía endosómica 0\n",
-      "['2', 'receptor'] 2 Receptor de manosa 2\n",
-      "['1', 'cd4'] 4 CD3 0\n",
-      "['4', 'lgm'] 3 lgG 0\n",
-      "['3'] 3 Transportando péptidos desde el citoplasma al retículo endoplásmico 1\n",
-      "['3'] 4 Hipermutación somática 0\n",
-      "['1', 'th1'] 1 Th1 2\n",
-      "['2', 'tcr'] 2 TCR de los linfocitos CD8 y CD4 2\n",
-      "['1', 'vcam'] 1 VCAM1 1\n",
-      "['1', 'c5'] 1 C5a 1\n",
-      "['2', 'durante'] 2 Durante su maduración y desarrollo en el timo 2\n",
-      "['3'] 3 Recombinación entre los segmentos VDJ reordenados y los genes constantes 1\n",
-      "['3', 'il'] 3 IL12 1\n",
-      "['1', 'cd16'] 1 CD16 FcRγIIIA 2\n",
-      "['2', 'il'] 2 IL10 y TGFβ 1\n",
-      "['1', 'lgg'] 1 lgG 2\n",
-      "['3', 'polis'] 3 Polisacáridos 1\n",
-      "['3', 'hla'] 3 HLADM 1\n",
-      "['2', 'c1'] 2 C1 inhibidor 2\n",
-      "['3', 'inhib'] 3 Inhibición de nuevos reordenamientos de las cadenas ligeras 1\n",
-      "['2', 'cd81'] 3 CD79α 0\n",
-      "['1', 'los', 'genes'] 1 Los genes variables de las inmunoglobulinas 3\n",
-      "['2', 'nk'] 2 NK 2\n",
-      "['3', 'cél'] 3 Células NK activadas por citocinas 1\n",
-      "['4', 'b'] 4 B 2\n",
-      "['4'] 4 Cuando la prueba cruzada es negativa 1\n",
-      "['3', 'cél'] 1 Células presentadoras de antígeno del donanteLinfocitos T del receptor 0\n",
-      "['3', 'bacter'] 3 Bacterias encapsuladas 1\n",
-      "['2', 'linf'] 1 Linfocitos Th1 0\n",
-      "['4', 'fish'] 4 FISH con sonda locusespecífica 2\n",
-      "['3'] 3 Utilizar material parafinado 1\n",
-      "['24'] 3 24 1\n",
-      "['3', '19'] 4 4p16 0\n",
-      "['1', '0'] 4 050 0\n",
-      "['2'] 4 GenBank 0\n",
-      "['2', 'ii'] 3 III 0\n",
-      "['1', '17'] 1 17 2\n",
-      "['2', 'lif'] 2 LiFraumeni 1\n",
-      "['4', 'explanation'] 4 Sano portador de una translocación robertsoniana 1\n",
-      "['1', '21'] 3 17q 0\n",
-      "['3', 'atm'] 3 ATM 2\n",
-      "['4', 'la', 'her'] 3 La varianza genética   aditiva por la varianza fenotípica 1\n",
-      "['2', 'atax'] 2 Ataxia de Friedrich 1\n",
-      "['2'] 2 Sirve como sitio de unión de los ribosomas para la traducción 1\n",
-      "['1', '9'] 4 151 0\n",
-      "['2', '52'] 2 52 2\n",
-      "['3', 'linf'] 3 Linfoma folicular 1\n",
-      "['3', 'ady'] 1 Adyacente tipo 2 0\n",
-      "['3', 'explanation'] 4 Por escisión de nucleótidos 0\n",
-      "['1', 'rubinstein'] 2 Alagille 0\n",
-      "['1', 'enfer'] 4 Síndrome de Rett 0\n",
-      "['3', 'transloc'] 3 Translocación 1\n",
-      "['2', 'se', 'puede'] 1 No se puede asumir que la variable sigue una distribución normal 2\n",
-      "['3', 'menor'] 1 Menor sea el nivel de confianza 1\n",
-      "['2', 'correl'] 2 Correlación de Pearson 1\n",
-      "['3', 'aden'] 3 Adenina timina citosina y guanina 1\n",
-      "['1', 'true'] 2 Las bases nitrogenadas se unen a la molécula de pentosa por su carbono 1 1\n",
-      "['3'] 4 Los grupos fosfato quedan expuestos hacia el exterior de la doble hélice 0\n",
-      "['1', '2'] 3 Para la corrección de los errores incorporados la polimerasa actúa en sentido 35 0\n",
-      "['3', 'dnapol'] 3 DNApolimerasa dependiente de RNA 1\n",
-      "['2', 'uaac'] 4 AAUGGCAAU 0\n",
-      "['3'] 1 El conjunto lineal de tripletes comprendido entre el codón de iniciación y el de terminación 0\n",
-      "['2', '3'] 2 Pueden contener más de un sitio de unión para el ribosoma 1\n",
-      "['4'] 4 La aparición de un codón stop en el sitio A del ribosoma 1\n",
-      "['1', 'cada'] 1 Cada codón codifica siempre el mismo aminoácido 2\n",
-      "['3'] 3 Secuencia de los cebadores 1\n",
-      "['1'] 4 Pueda replicarse autónomamente 0\n",
-      "['1', 'true'] 3 El clonaje en monocopia implica la integración del transgén en el cromosoma celular 0\n",
-      "['3', 'yac'] 3 YACs yeast artificial chromosomes 1\n",
-      "['3', 'a'] 1 A  T  G  C  1 1\n",
-      "['3', 'explanation'] 3 La metilación de una de las bases de la secuencia diana protege a ésta de la rotura por la endonucleasa correspondiente 1\n",
-      "['1'] 1 Detectar la transcripción de un gen en un organismo con una sonda de cDNA 1\n",
-      "['3', 'generar'] 1 Producir ATP por fosforilación a nivel de sustrato 0\n",
-      "['3', 'homol'] 3 Homoláctica 1\n",
-      "['3', 'σ'] 3 σ sigma 2\n",
-      "['1', 'una', 'elev'] 4 Asociaciones intermoleculares de 3 hélices extendidas 0\n",
-      "['4', 'explanation'] 1 No se une a la enzima en tanto no se haya formado el complejo enzimasustrato 0\n",
-      "['1'] 1 Un péptido natural con actividad reductora 1\n",
-      "['3', 'reducción'] 3 Reducción de los niveles de NADPH 2\n",
-      "['4', 'hipox'] 3 Ácido úrico 0\n",
-      "['2', 'citoc'] 2 Citocromo c oxidasa 1\n",
-      "['2', 'c'] 3 D 0\n",
-      "['2', 'prostag'] 4 Colesterol 0\n",
-      "['1', 'ure'] 2 Arginasa 0\n",
-      "['1', 's'] 1 SAdenosilmetionina 1\n",
-      "['1', 'glicer'] 1 Glicerol quinasa 1\n",
-      "['2', 'explanation'] 2 Facilita la formación del estado de transición 1\n",
-      "['2', 'repon'] 2 Reponer intermediarios del ciclo del ácido cítrico 1\n",
-      "['4', 'cobre'] 4 Cobre 2\n",
-      "['4'] 4 A veces consume ATP 1\n",
-      "['1', 'grupo', 'h'] 1 Grupo hemo 2\n",
-      "['2', 'los', 'organism'] 2 Los organismos ureotélicos 2\n",
-      "['1', 'hid'] 1 Hidroxilación de prolinas y lisinas en el colágeno 1\n",
-      "['3', 'notoc'] 3 Notocorda 1\n",
-      "['3', 'útero'] 4 Trompa uterina 0\n",
-      "['2', 'embar'] 1 Placenta previa 0\n",
-      "['3', 'ácido'] 3 Ácido periódicoreactivo de Schiff PAS 2\n",
-      "['2', 'fij'] 2 Fijación 1\n",
-      "['2', 'oví'] 1 Vivípara y tiene casi el doble de longitud que el macho 0\n",
-      "['1', 'hormig'] 1 Hormigas Formica fusca que contienen metacercarias 1\n",
-      "['2', 'anophe'] 2 Anopheles 1\n",
-      "['1'] 4 El cuerpo de la tenia 0\n",
-      "['2'] 2 Tienen aspecto de levaduras o de hongos filamentosos según las condiciones 1\n",
-      "['2', 'aldol'] 2 Aldolasa 1\n",
-      "['2', 'beta'] 4 Gamma 0\n",
-      "['4', 'circular', 'bic'] 1 Lineal bicatenario con una proteína unida a cada extremo 5 0\n",
-      "['1', 'falso'] 4 El complejo proteico entre ciclina y la cinasa dependiente de ciclina Cdk es un importante factor regulador del ciclo 0\n",
-      "['2', 'correl'] 3 Correlación de Spearman 0\n",
-      "['2', 'miast'] 2 Miastenia grave 1\n",
-      "['1', 'tej'] 1 Tejido conjuntivo 1\n",
-      "['1', 'vasop'] 1 Vasopresina o ADH 1\n",
-      "['4'] 3 5 0\n",
-      "ensemble test-tqa 0.738835725677831\n",
-      "ensemble train-tqa 0.7802846233946547\n",
-      "ensemble head-qa 0.6595744680851063\n",
-      "base test-tqa 0.7503987240829346\n",
-      "base train-tqa 0.7788962165914612\n",
-      "base head-qa 0.6978723404255319\n",
-      "base-base test-tqa 0.506872852233677\n",
-      "base-base head-qa 0.6140350877192983\n",
-      "single_traj test-tqa 0.7527910685805422\n",
-      "no_temp test-tqa 0.7615629984051037\n"
-     ]
-    }
-   ],
-   "source": [
-    "def lcs(x, y):\n",
-    "    m = len(x)\n",
-    "    n = len(y)\n",
-    "    dp = np.zeros((m + 1, n + 1), dtype=int)\n",
-    "\n",
-    "    for i in range(1, m + 1):\n",
-    "        for j in range(1, n + 1):\n",
-    "            if x[i - 1] == y[j - 1]:\n",
-    "                dp[i][j] = dp[i - 1][j - 1] + 1\n",
-    "            else:\n",
-    "                dp[i][j] = max(dp[i - 1][j], dp[i][j - 1])\n",
-    "\n",
-    "    return dp[m][n]\n",
-    "\n",
-    "\n",
-    "def measure(vals, verbose=False):\n",
-    "    accuracy = 0\n",
-    "\n",
-    "    correct_responses = []\n",
-    "    incorrect_responses = []\n",
-    "    for _, val in vals.items():\n",
-    "        val[\"expected_response\"] = re.sub(r\"[^\\w\\s]\", \"\", val[\"expected_response\"])\n",
-    "        val[\"response\"] = re.sub(r\"[^\\w\\s]\", \"\", val[\"response\"])\n",
-    "        hypothesis_words = val[\"response\"].lower().split()\n",
-    "        reference_words = val[\"expected_response\"].lower().split()\n",
-    "\n",
-    "        is_correct = lcs(hypothesis_words, reference_words)\n",
-    "\n",
-    "        if verbose:\n",
-    "            print(hypothesis_words, val[\"expected_response\"], is_correct)\n",
-    "        if is_correct:\n",
-    "            accuracy += 1\n",
-    "            correct_responses.append(val)\n",
-    "        else:\n",
-    "            incorrect_responses.append(val)\n",
-    "\n",
-    "    return {\"accuracy\": accuracy / len(vals), \"correct\": correct_responses, \"incorrect\": incorrect_responses}\n",
-    "\n",
-    "metrics = {}\n",
-    "for key, val in results.items():\n",
-    "    metric = {}\n",
-    "    for split, saved_vals in val.items():\n",
-    "        metric[split] = measure(saved_vals, verbose=key==\"base-base\")\n",
-    "    metrics[key] = metric\n",
-    "\n",
-    "for key, val in metrics.items():\n",
-    "    for keyb in metrics[key].keys():\n",
-    "        print(key, keyb, metrics[key][keyb][\"accuracy\"])"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 71,
-   "metadata": {},
-   "outputs": [
-    {
-     "data": {
-      "image/png": "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",
-      "text/plain": [
-       "<Figure size 640x480 with 1 Axes>"
-      ]
-     },
-     "metadata": {},
-     "output_type": "display_data"
-    },
-    {
-     "data": {
-      "image/png": "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",
-      "text/plain": [
-       "<Figure size 640x480 with 1 Axes>"
-      ]
-     },
-     "metadata": {},
-     "output_type": "display_data"
-    },
-    {
-     "data": {
-      "image/png": "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",
-      "text/plain": [
-       "<Figure size 640x480 with 1 Axes>"
-      ]
-     },
-     "metadata": {},
-     "output_type": "display_data"
-    },
-    {
-     "data": {
-      "image/png": "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",
-      "text/plain": [
-       "<Figure size 640x480 with 1 Axes>"
-      ]
-     },
-     "metadata": {},
-     "output_type": "display_data"
-    },
-    {
-     "data": {
-      "image/png": "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",
-      "text/plain": [
-       "<Figure size 640x480 with 1 Axes>"
-      ]
-     },
-     "metadata": {},
-     "output_type": "display_data"
-    }
-   ],
-   "source": [
-    "for model_type, d in metrics.items():\n",
-    "    if model_type not in [\"ensemble\", \"base\", \"base-base\", \"single_traj\", \"no_temp\"]:\n",
-    "        continue\n",
-    "\n",
-    "    plt.figure()\n",
-    "    for key in [\"test-tqa\", \"head-qa\"]:\n",
-    "        if key not in d:\n",
-    "            continue\n",
-    "\n",
-    "        for keyb in [\"correct\", \"incorrect\"]:\n",
-    "            responses = d[key][keyb]\n",
-    "    \n",
-    "            total_uncertainty = []\n",
-    "            for response in responses:\n",
-    "                total_uncertainty.append(np.average(response[\"total_uncertainty\"]))\n",
-    "\n",
-    "            plt.hist(total_uncertainty, bins=100, density=True, label=f\"{model_type}-{key}-{keyb}\", alpha=0.5)\n",
-    "\n",
-    "    plt.legend()\n",
-    "    plt.title(\"Total uncertainties on MCQ\")"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 70,
-   "metadata": {},
-   "outputs": [
-    {
-     "data": {
-      "text/plain": [
-       "<Figure size 640x480 with 0 Axes>"
-      ]
-     },
-     "metadata": {},
-     "output_type": "display_data"
-    },
-    {
-     "data": {
-      "image/png": "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",
-      "text/plain": [
-       "<Figure size 640x480 with 1 Axes>"
-      ]
-     },
-     "metadata": {},
-     "output_type": "display_data"
-    },
-    {
-     "data": {
-      "image/png": "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",
-      "text/plain": [
-       "<Figure size 640x480 with 1 Axes>"
-      ]
-     },
-     "metadata": {},
-     "output_type": "display_data"
-    }
-   ],
-   "source": [
-    "plt.figure()\n",
-    "\n",
-    "for model_type, d in metrics.items():\n",
-    "    if model_type not in [\"ensemble\", \"single_traj\"]:\n",
-    "        continue\n",
-    "\n",
-    "    plt.figure()\n",
-    "    for key in [\"test-tqa\", \"head-qa\"]:\n",
-    "        if key not in d:\n",
-    "            continue\n",
-    "\n",
-    "        for keyb in [\"correct\", \"incorrect\"]:\n",
-    "            responses = d[key][keyb]\n",
-    "    \n",
-    "            total_uncertainty = []\n",
-    "            for response in responses:\n",
-    "                total_uncertainty.append(np.average(response[\"aleatoric_uncertainty\"]))\n",
-    "\n",
-    "            plt.hist(total_uncertainty, bins=100, density=True, label=f\"{model_type}-{key}-{keyb}\", alpha=0.5)\n",
-    "\n",
-    "    plt.legend()\n",
-    "    plt.title(\"Epistemic uncertainties on MCQ\")"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": []
-  }
- ],
- "metadata": {
-  "kernelspec": {
-   "display_name": "posteriors",
-   "language": "python",
-   "name": "python3"
-  },
-  "language_info": {
-   "codemirror_mode": {
-    "name": "ipython",
-    "version": 3
-   },
-   "file_extension": ".py",
-   "mimetype": "text/x-python",
-   "name": "python",
-   "nbconvert_exporter": "python",
-   "pygments_lexer": "ipython3",
-   "version": "3.11.9"
-  }
- },
- "nbformat": 4,
- "nbformat_minor": 2
-}
diff --git a/examples/bayes_llama3/data_explore.ipynb b/examples/bayes_llama3/data_explore.ipynb
deleted file mode 100644
index 2bcbe156..00000000
--- a/examples/bayes_llama3/data_explore.ipynb
+++ /dev/null
@@ -1,3438 +0,0 @@
-{
- "cells": [
-  {
-   "cell_type": "code",
-   "execution_count": 1,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "import json, os\n",
-    "\n",
-    "with open(\"/home/paperspace/Projects/posteriors/datasets/tqa_train_val_test/test/tqa_v2_test.json\") as f:\n",
-    "    qa_doc = json.load(f)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 5,
-   "metadata": {},
-   "outputs": [
-    {
-     "data": {
-      "text/plain": [
-       "{'T_0271': {'content': {'figures': [{'caption': 'FIGURE 16.11 A coast guard officer looks for survivors of Hurricane Katrina.',\n",
-       "     'imagePath': 'textbook_images/storms_20166.png'}],\n",
-       "   'mediaLinks': [],\n",
-       "   'text': 'A storm is an episode of severe weather caused by a major disturbance in the atmosphere. Storms can vary a lot in the time they last and in how severe they are. A storm may last for less than an hour or for more than a week. It may affect just a few square kilometers or thousands. Some storms are harmless and some are disastrous. The size and strength of a storm depends on the amount of energy in the atmosphere. Greater differences in temperature and air pressure produce stronger storms. Types of storms include thunderstorms, tornadoes, hurricanes, and winter storms such as blizzards. '},\n",
-       "  'globalID': 'T_0271',\n",
-       "  'topicName': 'What Are Storms'},\n",
-       " 'T_0272': {'content': {'figures': [],\n",
-       "   'mediaLinks': [],\n",
-       "   'text': 'Thunderstorms are are known for their heavy rains and lightning. In strong thunderstorms, hail and high winds are also likely. Thunderstorms are very common. Worldwide, there are about 14 million of them each year! In the U.S., they are most common and strongest in the Midwest. '},\n",
-       "  'globalID': 'T_0272',\n",
-       "  'topicName': 'Thunderstorms'},\n",
-       " 'T_0273': {'content': {'figures': [{'caption': 'FIGURE 16.12 A thunderhead is a cumulonimbus cloud.',\n",
-       "     'imagePath': 'textbook_images/storms_20167.png'}],\n",
-       "   'mediaLinks': [],\n",
-       "   'text': 'Thunderstorms occur when the air is very warm and humid. The warm air rises rapidly to create strong updrafts. When the rising air cools, its water vapor condenses. The updrafts create tall cumulonimbus clouds called thunder- heads. You can see one in Figure 16.12. '},\n",
-       "  'globalID': 'T_0273',\n",
-       "  'topicName': 'What Causes Thunderstorms'},\n",
-       " 'T_0274': {'content': {'figures': [{'caption': 'FIGURE 16.13 Lightning flashes across an Arizona sun- set.',\n",
-       "     'imagePath': 'textbook_images/storms_20168.png'}],\n",
-       "   'mediaLinks': [],\n",
-       "   'text': 'During a thunderstorm, some parts of a thunderhead become negatively charged. Other parts become positively charged. The difference in charge creates lightning. Lightning is a huge release of electricity. Lightning can jump between oppositely charged parts of the same cloud, between one cloud and another, or between a cloud and the ground. You can see lightning in Figure 16.13. Lightning blasts the air with energy. The air heats and expands so quickly that it explodes. This creates the loud sound of thunder. Do you know why you always hear the boom of thunder after you see the flash of lightning? Its because light travels faster than sound. If you count the seconds between seeing lightning and hearing thunder, you can estimate how far away the lightning was. A lapse of 5 seconds is equal to about a mile. '},\n",
-       "  'globalID': 'T_0274',\n",
-       "  'topicName': 'Lightning and Thunder'},\n",
-       " 'T_0275': {'content': {'figures': [],\n",
-       "   'mediaLinks': [],\n",
-       "   'text': 'Severe thunderstorms have a lot of energy and strong winds. This allows them to produce tornadoes. A tornado is a funnel-shaped cloud of whirling high winds. You can see a tornado in Figure 16.14. The funnel moves along the ground, destroying everything in its path. As it moves it loses energy. Before this happens it may have gone up to 25 kilometers (16 miles). Fortunately, tornadoes are narrow. They may be only 150 meters (500 feet) wide. '},\n",
-       "  'globalID': 'T_0275',\n",
-       "  'topicName': 'Tornadoes'},\n",
-       " 'T_0276': {'content': {'figures': [{'caption': 'FIGURE 16.14 Tornadoes are small but mighty storms.',\n",
-       "     'imagePath': 'textbook_images/storms_20169.png'}],\n",
-       "   'mediaLinks': [],\n",
-       "   'text': 'The winds of a tornado can reach very high speeds. The faster the winds blow, the greater the damage they cause. Wind speed and damage are used to classify tornadoes. Table 16.1 shows how. F Scale F0 (km/hr) 64-116 (mph) 40-72 F1 117-180 73-112 Damage Light - tree branches fall and chimneys may col- lapse Moderate - mobile homes, autos pushed aside F Scale F2 (km/hr) 181-253 (mph) 113-157 F3 254-332 158-206 F4 333-419 207-260 F5 420-512 261-318 F6 >512 >318 Damage Considerable - roofs torn off houses, large trees up- rooted Severe - houses torn apart, trees uprooted, cars lifted Devastating - houses lev- eled, cars thrown Incredible - structures fly, cars become missiles Maximum tornado wind speed '},\n",
-       "  'globalID': 'T_0276',\n",
-       "  'topicName': 'Classifying Tornadoes'},\n",
-       " 'T_0277': {'content': {'figures': [{'caption': 'FIGURE 16.15 Tornadoes are most common in the cen- tral part of the U.S.',\n",
-       "     'imagePath': 'textbook_images/storms_20170.png'}],\n",
-       "   'mediaLinks': [],\n",
-       "   'text': 'Look at the map in Figure 16.15. It shows where the greatest number of tornadoes occur in the U.S. Tornadoes can happen almost anywhere in the U.S. but only this area is called tornado alley. Why do so many tornadoes occur here? This is where warm air masses from the south run into cold air masses from the north. '},\n",
-       "  'globalID': 'T_0277',\n",
-       "  'topicName': 'Tornado Alley'},\n",
-       " 'T_0278': {'content': {'figures': [],\n",
-       "   'mediaLinks': [],\n",
-       "   'text': 'Tornadoes may also come from hurricanes. A hurricane is an enormous storm with high winds and heavy rains. Hurricanes may be hundreds of kilometers wide. They may travel for thousands of kilometers. The storms wind speeds may be greater than 251 kilometers (156 miles) per hour. Hurricanes develop from tropical cyclones. Hurricanes form over warm very ocean water. This water gives them their energy. As long as a hurricane stays over the warm ocean, it keeps growing stronger. However, if it goes ashore or moves over cooler water, it is cut off from the hot water energy. The storm then loses strength and slowly fades away. '},\n",
-       "  'globalID': 'T_0278',\n",
-       "  'topicName': 'Hurricanes'},\n",
-       " 'T_0279': {'content': {'figures': [{'caption': 'FIGURE 16.16 The eye of this hurricane is easy to see from space.',\n",
-       "     'imagePath': 'textbook_images/storms_20171.png'}],\n",
-       "   'mediaLinks': [],\n",
-       "   'text': 'At the center of a hurricane is a small area where the air is calm and clear. This is the eye of the hurricane. The eye forms at the low-pressure center of the hurricane. You can see the eye of a hurricane in Figure 16.16. '},\n",
-       "  'globalID': 'T_0279',\n",
-       "  'topicName': 'The Eye of a Hurricane'},\n",
-       " 'T_0280': {'content': {'figures': [],\n",
-       "   'mediaLinks': [],\n",
-       "   'text': 'Like tornadoes, hurricanes are classified on the basis of wind speed and damage. Table 16.2 shows how. Category 1 (weak) Kph 119-153 Mph 74-95 2 (moderate) 154-177 96-110 3 (strong) 178-209 111-130 Damage Above normal; no real damage to structures Some roofing, door, and window damage, consid- erable damage to vegeta- tion, mobile homes, and piers Some buildings damaged; mobile homes destroyed Category 4 (very strong) Kph 210-251 Mph 131-156 5 (devastating) >251 >156 Damage Complete roof failure on small residences; major erosion of beach areas; major damage to lower floors of structures near shore Complete roof failure on many residences and in- dustrial buildings; some complete building failures '},\n",
-       "  'globalID': 'T_0280',\n",
-       "  'topicName': 'Classifying Hurricanes'},\n",
-       " 'T_0281': {'content': {'figures': [{'caption': 'FIGURE 16.17 Storm surge can cause serious flooding.',\n",
-       "     'imagePath': 'textbook_images/storms_20172.png'}],\n",
-       "   'mediaLinks': [],\n",
-       "   'text': 'Some of the damage from a hurricane is caused by storm surge. Storm surge is very high water located in the low pressure eye of the hurricane. The very low pressure of the eye allows the water level to rise above normal sea level. Storm surge can cause flooding when it reaches land. You can see this in Figure 16.17. High winds do a great deal of damage in hurricanes. High winds can also create very big waves. If the large waves are atop a storm surge, the high water can flood the shore. If the storm happens to occur at high tide, the water will rise even higher. '},\n",
-       "  'globalID': 'T_0281',\n",
-       "  'topicName': 'Storm Surge'},\n",
-       " 'T_0282': {'content': {'figures': [],\n",
-       "   'mediaLinks': [],\n",
-       "   'text': 'Like hurricanes, winter storms develop from cyclones. But in the case of winter storms, the cyclones form at higher latitudes. In North America, cyclones often form when the jet stream dips south in the winter. This lets dry polar air pour south. At the same time, warm moist air from the Gulf of Mexico flows north. When the two air masses meet, the differences in temperature and pressure cause strong winds and heavy precipitation. Two types of winter storms that occur in the U.S. are blizzards and lake-effect snow storms. '},\n",
-       "  'globalID': 'T_0282',\n",
-       "  'topicName': 'Winter Storms'},\n",
-       " 'T_0283': {'content': {'figures': [{'caption': 'FIGURE 16.18 Blizzard in Washington, D.C. Blizzards are unusual in Washington, D.C many parts of the United States. Do they ever occur where you live?',\n",
-       "     'imagePath': 'textbook_images/storms_20173.png'}],\n",
-       "   'mediaLinks': [],\n",
-       "   'text': 'A blizzard is a snow storm that has high winds. To be called a blizzard, a storm must have winds greater than 56 kilometers (35 miles) per hour and visibility of 14 mile or less because of wind-blown snow. You can see a blizzard in Figure 16.18. Blizzards are dangerous storms. The wind may blow the snow into deep drifts. Along with the poor visibility, the snow drifts make driving risky. The wind also makes cold temperatures more dangerous. The greater the wind speed, the higher the windchill. Windchill is what the temperature feels like when the wind is taken into account. It depends on air temperature and wind speed, as you can see in Figure 16.19. Higher windchill will cause a person to suffer frostbite and other harmful effects of cold sooner than if the wind isnt blowing. '},\n",
-       "  'globalID': 'T_0283',\n",
-       "  'topicName': 'Blizzards'},\n",
-       " 'T_0284': {'content': {'figures': [],\n",
-       "   'mediaLinks': [],\n",
-       "   'text': 'Some places receive very heavy snowfall just about every winter. If they are near a lake, they may be getting lake- effect snow. Figure 16.20 shows how lake-effect snow occurs. Winter winds pick up moisture as they pass over the relatively warm waters of a large lake. When the winds reach the cold land on the other side, the air cools. Since there was so much moisture in the air it can drop a lot of snow. More than 254 centimeters (100 inches) of snow may fall in a single lake-effect storm! '},\n",
-       "  'globalID': 'T_0284',\n",
-       "  'topicName': 'LakeEffect Snow'}}"
-      ]
-     },
-     "execution_count": 5,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "qa_doc[0].keys()\n",
-    "qa_doc[5][\"topics\"]"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 3,
-   "metadata": {},
-   "outputs": [
-    {
-     "data": {
-      "text/plain": [
-       "{'advances in genetics',\n",
-       " 'air masses',\n",
-       " 'air movement',\n",
-       " 'air pressure and altitude',\n",
-       " 'alligators and crocodiles',\n",
-       " 'amontonss law',\n",
-       " 'amphibians',\n",
-       " 'angiosperms',\n",
-       " 'animal behaviors',\n",
-       " 'animal communication',\n",
-       " 'animal like protists',\n",
-       " 'arachnids',\n",
-       " 'archaea',\n",
-       " 'arthropods',\n",
-       " 'basic and applied science',\n",
-       " 'behavior of gases',\n",
-       " 'bernoullis law',\n",
-       " 'biodiversity and extinction',\n",
-       " 'biological communities',\n",
-       " 'biotechnology in agriculture',\n",
-       " 'bird reproduction',\n",
-       " 'birds',\n",
-       " 'blizzards',\n",
-       " 'blood pressure',\n",
-       " 'boyles law',\n",
-       " 'branches of earth science',\n",
-       " 'centipedes and millipedes',\n",
-       " 'changing weather',\n",
-       " 'charless law',\n",
-       " 'choosing healthy foods',\n",
-       " 'chordates',\n",
-       " 'climate and its causes',\n",
-       " 'cnidarians',\n",
-       " 'collecting weather data',\n",
-       " 'communication in science',\n",
-       " 'communities',\n",
-       " 'competition',\n",
-       " 'consumers and decomposers',\n",
-       " 'control of insects',\n",
-       " 'crustaceans',\n",
-       " 'cyclic behavior of animals',\n",
-       " 'development of theories',\n",
-       " 'diversity of birds',\n",
-       " 'echinoderms and invertebrate chordates',\n",
-       " 'ecosystem change',\n",
-       " 'ecosystems',\n",
-       " 'effect of continental position on climate',\n",
-       " 'effects of air pollution on the environment',\n",
-       " 'erosion and deposition by wind',\n",
-       " 'ethics in science',\n",
-       " 'evolution and classification of plants',\n",
-       " 'evolution plate tectonics and climate change',\n",
-       " 'extinction and radiation of life',\n",
-       " 'field study',\n",
-       " 'fields in the life sciences',\n",
-       " 'flow of energy',\n",
-       " 'flow of matter in ecosystems',\n",
-       " 'food webs',\n",
-       " 'frogs and toads',\n",
-       " 'fungi',\n",
-       " 'fungi classification',\n",
-       " 'fungi reproduction',\n",
-       " 'gases',\n",
-       " 'global wind belts',\n",
-       " 'gymnosperms',\n",
-       " 'habitat and niche',\n",
-       " 'habitat destruction',\n",
-       " 'history of cenozoic life',\n",
-       " 'history of earths life forms',\n",
-       " 'history of mesozoic life',\n",
-       " 'history of paleozoic life',\n",
-       " 'history of science',\n",
-       " 'how animals evolved',\n",
-       " 'human uses of fungi',\n",
-       " 'human vision',\n",
-       " 'humans and primates',\n",
-       " 'hurricanes',\n",
-       " 'hypothesis',\n",
-       " 'importance of arthropods',\n",
-       " 'importance of biodiversity',\n",
-       " 'importance of birds',\n",
-       " 'importance of echinoderms',\n",
-       " 'importance of insects',\n",
-       " 'importance of mammals',\n",
-       " 'importance of mollusks',\n",
-       " 'importance of protists',\n",
-       " 'importance of reptiles',\n",
-       " 'importance of seedless plants',\n",
-       " 'innate behavior of animals',\n",
-       " 'insect food',\n",
-       " 'insect reproduction and life cycle',\n",
-       " 'insects',\n",
-       " 'insects and other arthropods',\n",
-       " 'introduction to ecology',\n",
-       " 'introduction to genetics',\n",
-       " 'introduction to plants',\n",
-       " 'invertebrates',\n",
-       " 'learned behavior of animals',\n",
-       " 'levels of ecological organization',\n",
-       " 'lizards and snakes',\n",
-       " 'local winds',\n",
-       " 'mammal characteristics',\n",
-       " 'mammal classification',\n",
-       " 'mammal reproduction',\n",
-       " 'mammals',\n",
-       " 'mass extinctions',\n",
-       " 'mendels discoveries',\n",
-       " 'mendels laws and genetics',\n",
-       " 'mendels pea plants',\n",
-       " 'microevolution and macroevolution',\n",
-       " 'mid latitude cyclones',\n",
-       " 'modern biodiversity',\n",
-       " 'modern genetics',\n",
-       " 'molecular evidence for evolution',\n",
-       " 'natural selection',\n",
-       " 'nature of science',\n",
-       " 'nonvascular plants',\n",
-       " 'observation',\n",
-       " 'observations and experiments',\n",
-       " 'oceanic pressure',\n",
-       " 'organization of living things',\n",
-       " 'origin of species',\n",
-       " 'pascals law',\n",
-       " 'physical science careers',\n",
-       " 'plant characteristics',\n",
-       " 'plant classification',\n",
-       " 'plant hormones',\n",
-       " 'plant like protists',\n",
-       " 'plant responses and special adaptations',\n",
-       " 'plants adaptations for life on land',\n",
-       " 'predation',\n",
-       " 'predicting weather',\n",
-       " 'pressure and density of the atmosphere',\n",
-       " 'pressure in fluids',\n",
-       " 'pressure of fluids',\n",
-       " 'primates',\n",
-       " 'protist characteristics',\n",
-       " 'protists nutrition',\n",
-       " 'punnett squares',\n",
-       " 'replication in science',\n",
-       " 'reproduction in seedless plants',\n",
-       " 'reproductive behavior of animals',\n",
-       " 'reptiles',\n",
-       " 'role of amphibians',\n",
-       " 'roles in an ecosystem',\n",
-       " 'safety in life science research',\n",
-       " 'safety in science',\n",
-       " 'safety in the life sciences',\n",
-       " 'salamanders',\n",
-       " 'science skills',\n",
-       " 'scientific community',\n",
-       " 'scientific experiments',\n",
-       " 'scientific explanations and interpretations',\n",
-       " 'scientific induction',\n",
-       " 'scientific investigation',\n",
-       " 'scientific law',\n",
-       " 'scientific method',\n",
-       " 'scientific process',\n",
-       " 'scientific theories',\n",
-       " 'scientific theory',\n",
-       " 'scientific ways of thinking',\n",
-       " 'scope of physical science',\n",
-       " 'seasonal changes in plants',\n",
-       " 'seeds and seed dispersal',\n",
-       " 'social behavior of animals',\n",
-       " 'storms',\n",
-       " 'structural evidence for evolution',\n",
-       " 'succession',\n",
-       " 'symbiosis',\n",
-       " 'symbiotic relationships of fungi',\n",
-       " 'technological design process',\n",
-       " 'technology',\n",
-       " 'technology and science',\n",
-       " 'technology careers',\n",
-       " 'temperature of the atmosphere',\n",
-       " 'terrestrial biomes',\n",
-       " 'the biosphere',\n",
-       " 'the nature of science',\n",
-       " 'the scientific method',\n",
-       " 'the scope of physical science',\n",
-       " 'tornadoes',\n",
-       " 'tracing evolution',\n",
-       " 'tropisms',\n",
-       " 'turtles',\n",
-       " 'types of animal behavior',\n",
-       " 'types of archaea',\n",
-       " 'types of marine organisms',\n",
-       " 'types of mollusks',\n",
-       " 'understanding animal behavior',\n",
-       " 'vascular seedless plants',\n",
-       " 'vertebrate characteristics',\n",
-       " 'weather forecasting',\n",
-       " 'weather fronts',\n",
-       " 'weather maps',\n",
-       " 'weather versus climate',\n",
-       " 'what are animals',\n",
-       " 'what are biomes',\n",
-       " 'what is life science',\n",
-       " 'what is science',\n",
-       " 'wind power',\n",
-       " 'women and people of color in science'}"
-      ]
-     },
-     "execution_count": 3,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "from collections import Counter\n",
-    "counter = Counter([d[\"lessonName\"] for d in qa_doc])\n",
-    "set(counter.keys())"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 29,
-   "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "the nature of science\n",
-      "T_0001 The scientific method is a set of steps that help us to answer questions. When we use logical steps and control the number of things that can be changed, we get better answers. As we test our ideas, we may come up with more questions. The basic sequence of steps followed in the scientific method is illustrated in Figure 1.1. \n",
-      "T_0002 Asking a question is one really good way to begin to learn about the natural world. You might have seen something that makes you curious. You might want to know what to change to produce a better result. Lets say a farmer is having an erosion problem. She wants to keep more soil on her farm. The farmer learns that a farming method called no-till farming allows farmers to plant seeds without plowing the land. She wonders if planting seeds without plowing will reduce the erosion problem and help keep more soil on her farmland. Her question is this: Will using the no-till method of farming help me to lose less soil on my farm? (Figure 1.2). \n",
-      "T_0003 Before she begins, the farmer needs to learn more about this farming method. She can look up information in books and magazines in the library. She may also search the Internet. A good way for her to learn is to talk to people who have tried this way of farming. She can use all of this information to figure out how she is going to test her question about no-till farming. Farming machines are shown in the Figure 1.3. \n",
-      "T_0004 After doing the research, the farmer will try to answer the question. She might think, If I dont plow my fields, I will lose less soil than if I do plow the fields. Plowing disrupts the soil and breaks up roots that help hold soil in place. This answer to her question is a hypothesis. A hypothesis is a reasonable explanation. A hypothesis can be tested. It may be the right answer, it may be a wrong answer, but it must be testable. Once she has a hypothesis, the next step is to do experiments to test the hypothesis. A hypothesis can be proved or disproved by testing. If a hypothesis is repeatedly tested and shown to be true, then scientists call it a theory. \n",
-      "T_0005 When we design experiments, we choose just one thing to change. The thing we change is called the independent variable. In the example, the farmer chooses two fields and then changes only one thing between them. She changes how she plows her fields. One field will be tilled and one will not. Everything else will be the same on both fields: the type of crop she grows, the amount of water and fertilizer that she uses, and the slope of the fields she plants on. The fields should be facing the same direction to get about the same amount of sunlight. These are the experimental controls. If the farmer only changes how she plows her fields, she can see the impact of the one change. After the experiment is complete, scientists then measure the result. The farmer measures how much soil is lost from each field. This is the dependent variable. How much soil is lost from each field depends on the plowing method. \n",
-      "T_0006 During an experiment, a scientist collects data. The data might be measurements, like the farmer is taking in Figure labeled. Labeling helps the scientist to know what each number represents. A scientist may also write descriptions of what happened during the experiment. At the end of the experiment the scientist studies the data. The scientist may create a graph or drawing to show the data. If the scientist can picture the data the results may be easier to understand. Then it is easier to draw logical conclusions. Even if the scientist is really careful it is possible to make a mistake. One kind of mistake is with the equipment. For example, an electronic balance may always measure one gram high. To fix this, the balance should be adjusted. If it cant be adjusted, each measurement should be corrected. A mistake can come if a measurement is hard to make. For example, the scientist may stop a stopwatch too soon or too late. To fix this, the scientist should run the experiment many times and make many measurements. The average of the measurements will be the accurate answer. Sometimes the result from one experiment is very different from the other results. If one data point is really different, it may be thrown out. It is likely a mistake was made in that experiment. \n",
-      "T_0007 The scientist must next form a conclusion. The scientist must study all of the data. What statement best explains the data? Did the experiment prove the hypothesis? Sometimes an experiment shows that a hypothesis is correct. Other times the data disproves the hypothesis. Sometimes its not possible to tell. If there is no conclusion, the scientist may test the hypothesis again. This time he will use some different experiments. No matter what the experiment shows the scientist has learned something. Even a disproved hypothesis can lead to new questions. The farmer grows crops on the two fields for a season. She finds that 2.2 times as much soil was lost on the plowed field as compared to the unplowed field. She concludes that her hypothesis was correct. The farmer also notices some other differences in the two plots. The plants in the no-till plots are taller. The soil moisture seems higher. She decides to repeat the experiment. This time she will measure soil moisture, plant growth, and the total amount of water the plants consume. From now on she will use no-till methods of farming. She will also research other factors that may reduce soil erosion. \n",
-      "T_0008 When scientists have the data and conclusions, they write a paper. They publish their paper in a scientific journal. A journal is a magazine for the scientists who are interested in a certain field. Before the paper is printed, other scientists look at it to try to find mistakes. They see if the conclusions follow from the data. This is called peer review. If the paper is sound it is printed in the journal. Other papers are published on the same topic in the journal. The evidence for or against a hypothesis is discussed by many scientists. Sometimes a hypothesis is repeatedly shown to be true and never shown to be false. The hypothesis then becomes a theory. Sometimes people say they have a theory when what they have is a hypothesis. In science, a theory has been repeatedly shown to be true. A theory is supported by many observations. However, a theory may be disproved if conflicting data is discovered. Many important theories have been shown to be true by many observations and experiments and are extremely unlikely to be disproved. These include the theory of plate tectonics and the theory of evolution. \n",
-      "T_0009 Scientists use models to help them understand and explain ideas. Models explain objects or systems in a more simple way. Models often only show only a part of a system. The real situation is more complicated. Models help scientists to make predictions about complex systems. Some models are something that you can see or touch. Other types of models use an idea or numbers. Each type is useful in certain ways. Scientists create models with computers. Computers can handle enormous amounts of data. This can more accu- rately represent the real situation. For example, Earths climate depends on an enormous number of factors. Climate models can predict how climate will change as certain gases are added to the atmosphere. To test how good a model is, scientists might start a test run at a time in the past. If the model can predict the present it is probably a good model. It is more likely to be accurate when predicting the future. \n",
-      "T_0010 A physical model is a representation of something using objects. It can be three-dimensional, like a globe. It can also be a two-dimensional drawing or diagram. Models are usually smaller and simpler than the real object. They most likely leave out some parts, but contain the important parts. In a good model the parts are made or drawn to scale. Physical models allow us to see, feel and move their parts. This allows us to better understand the real system. An example of a physical model is a drawing of the layers of Earth (Figure 1.5). A drawing helps us to understand the structure of the planet. Yet there are many differences between a drawing and the real thing. The size of a model is much smaller, for example. A drawing also doesnt give good idea of how substances move. Arrows showing the direction the material moves can help. A physical model is very useful but it cant explain the real Earth perfectly. \n",
-      "T_0011 Some models are based on an idea that helps scientists explain something. A good idea explains all the known facts. An example is how Earth got its Moon. A Mars-sized planet hit Earth and rocky material broke off of both bodies (Figure 1.6). This material orbited Earth and then came together to form the Moon. This is a model of something that happened billions of years ago. It brings together many facts known from our studies of the Moons surface. It accounts for the chemical makeup of rocks from the Moon, Earth, and meteorites. The physical properties of Earth and Moon figure in as well. Not all known data fits this model, but much does. There is also more information that we simply dont yet know. \n",
-      "T_0012 Models may use formulas or equations to describe something. Sometimes math may be the only way to describe it. For example, equations help scientists to explain what happened in the early days of the universe. The universe formed so long ago that math is the only way to describe it. A climate model includes lots of numbers, including temperature readings, ice density, snowfall levels, and humidity. These numbers are put into equations to make a model. The results are used to predict future climate. For example, if there are more clouds, does global temperature go up or down? Models are not perfect because they are simple versions of the real situation. Even so, these models are very useful to scientists. These days, models of complex things are made on computers. \n",
-      "T_0013 Accidents happen from time to time in everyday life. Since science involves an adventure into the unknown, it is natural that accidents can happen. Therefore, we must be careful and use proper equipment to prevent accidents (Figure 1.7). We must also be sure to treat any injury or accident appropriately. \n",
-      "T_0014 If you work in the science lab, you may come across dangerous materials or situations. Sharp objects, chemicals, heat, and electricity are all used at times in science laboratories. With proper protection and precautions, almost all accidents can be prevented (Figure 1.8). If an accident happens, it can be dealt with appropriately. Below is a list of safety guidelines to follow when doing labs: Follow directions at all times. A science lab is not a play area. Be sure to obey all safety guidelines given in lab instructions or by the lab supervisor. Be sure to use the correct amount of each material. Tie back long hair. Wear closed shoes with flat heels. Shirts should have no hanging sleeves, hoods, or drawstrings. Use gloves, goggles, or safety aprons as instructed. Be very careful when you use sharp or pointed objects, such as knives. Clean up broken glass quickly with a dust pan and broom. Never touch broken glass with your bare hands. Never eat or drink in the science lab. Table tops and counters could have dangerous substances on them. Keep your work area neat and clean. A messy work area can lead to spills and breakage. Completely clean materials like test tubes and beakers. Leftover substances could interact with other sub- stances in future experiments. If you are using flames or heat plates, be careful when you reach. Be sure your arms and hair are kept far away from heat sources. Use electrical appliances and burners as instructed. Know how to use an eye wash station, fire blanket, fire extinguisher, and first aid kit. Alert the lab supervisor if anything unusual occurs. Fill out an accident report if someone is hurt. The lab supervisor must know if any materials are damaged or discarded. \n",
-      "T_0015 Many Earth science investigations are conducted in the field (Figure 1.9). Field work needs some additional precautions: Be sure to wear appropriate clothing. Hiking requires boots, long pants, and protection from the Sun, for example. Bring sufficient supplies like food and water, even for a short trip. Dehydration can occur rapidly. Take along first aid supplies. Let others know where you are going, what you will be doing, and when you will be returning. Take a map with you if you dont know the area and leave a copy of the map with someone at home. Try to have access to emergency services and some way to communicate. Beware that cell phones may not have coverage in all locations. Be sure that you are accompanied by a person familiar with the area or is familiar with field work. \n",
-      "\n",
-      "NDQ_000001\n",
-      "Steps of the scientific method include all of the following except\n",
-      "a. doing background research.\n",
-      "b. constructing a hypothesis.\n",
-      "c. asking a question.\n",
-      "d. proving a theory.\n",
-      "d. proving a theory.\n",
-      "NDQ_000002\n",
-      "Why do scientists call the Big Bang a theory?\n",
-      "a. It is probably unlikely and therefore not a fact.\n",
-      "b. A very well respected scientist proved it to be true.\n",
-      "c. Many scientists have agreed upon this explanation after repeated experiments and models have shown it\n",
-      "d. All possible answers to a scientific idea are called theories.\n",
-      "c. Many scientists have agreed upon this explanation after repeated experiments and models have shown it\n",
-      "NDQ_000003\n",
-      "The data collected in an experiment should always be\n",
-      "a. labeled.\n",
-      "b. recorded.\n",
-      "c. reported.\n",
-      "d. all of the above\n",
-      "d. all of the above\n",
-      "NDQ_000004\n",
-      "Which of the following is not a scientific model?\n",
-      "a. A cross section of an apple that mimics the layers of the Earth.\n",
-      "b. A chart with nutritional information about food we eat.\n",
-      "c. A computer simulation that can show what will happen to algae in a pond over 10 years given conditions\n",
-      "d. An explanation for the extinction of the dinosaurs that takes into account volcanic activity, climate, space\n",
-      "b. A chart with nutritional information about food we eat.\n",
-      "NDQ_000005\n",
-      "If the results of an experiment disprove a hypothesis, then the\n",
-      "a. results should not be reported.\n",
-      "b. hypothesis is just a theory.\n",
-      "c. data must contain errors.\n",
-      "d. none of the above\n",
-      "d. none of the above\n",
-      "NDQ_000006\n",
-      "Which of the following are good measures to follow when working in the field?\n",
-      "a. Bring sun protection and sufficient water.\n",
-      "b. Do not travel without someone who knows the area.\n",
-      "c. Bring first aid supplies.\n",
-      "d. More than one answer is correct.\n",
-      "d. More than one answer is correct.\n",
-      "NDQ_000007\n",
-      "A scientist is conducting an experiment to determine which of three building structure types will best withstand the force of an earthquake. Which of the following is most likely to be the dependent variable?\n",
-      "a. The amount of damage each building receives.\n",
-      "b. The magnitude of the earthquake.\n",
-      "c. The structure of the building.\n",
-      "d. The type of soil each building is sitting on\n",
-      "a. The amount of damage each building receives.\n",
-      "NDQ_000008\n",
-      "Which statement about a scientific theory is false?\n",
-      "a. A theory can never be disproven.\n",
-      "b. A theory is supported by many observations.\n",
-      "c. A theory may develop from a well-supported hypothesis.\n",
-      "d. A theory may be rejected if conflicting data are discovere\n",
-      "a. A theory can never be disproven.\n",
-      "NDQ_000009\n",
-      "Types of scientific models include\n",
-      "a. mathematical equations.\n",
-      "b. computer models.\n",
-      "c. physical models.\n",
-      "d. all of the above\n",
-      "d. all of the above\n",
-      "NDQ_000010\n",
-      "Conclusions in an experiment\n",
-      "a. Improve with greater and more accurate dat.\n",
-      "b. Often lead a researcher to new scientific questions\n",
-      "c. Can agree or disagree with the hypothesis.\n",
-      "d. All of the above.\n",
-      "d. All of the above.\n",
-      "NDQ_000012\n",
-      "Which of the following is a lab safety rule?\n",
-      "a. You may drink but not eat in the lab.\n",
-      "b. You should tie back your hair if it is long.\n",
-      "c. You may wear sandals but not flip-flops in the lab.\n",
-      "d. You should leave used glassware for your teacher to wash.\n",
-      "b. You should tie back your hair if it is long.\n",
-      "NDQ_000014\n",
-      "representation of something using objects\n",
-      "a. control\n",
-      "b. dependent variable\n",
-      "c. hypothesis\n",
-      "d. independent variable\n",
-      "e. physical model\n",
-      "f. theory\n",
-      "g. scientific method\n",
-      "e. physical model\n",
-      "NDQ_000017\n",
-      "factor that is held constant in a scientific experiment\n",
-      "a. control\n",
-      "b. dependent variable\n",
-      "c. hypothesis\n",
-      "d. independent variable\n",
-      "e. physical model\n",
-      "f. theory\n",
-      "g. scientific method\n",
-      "a. control\n",
-      "NDQ_000019\n",
-      "variable that is changed in an experiment to see how it affects another variable\n",
-      "a. control\n",
-      "b. dependent variable\n",
-      "c. hypothesis\n",
-      "d. independent variable\n",
-      "e. physical model\n",
-      "f. theory\n",
-      "g. scientific method\n",
-      "d. independent variable\n",
-      "NDQ_000020\n",
-      "The scientific method is used to answer any question that one can think of.\n",
-      "a. true\n",
-      "b. false\n",
-      "b. false\n",
-      "NDQ_000021\n",
-      "scientific explanation that is widely accepted because it has been tested repeatedly and not proven false\n",
-      "a. control\n",
-      "b. dependent variable\n",
-      "c. hypothesis\n",
-      "d. independent variable\n",
-      "e. physical model\n",
-      "f. theory\n",
-      "g. scientific method\n",
-      "f. theory\n",
-      "NDQ_000022\n",
-      "Scientific models are an organized step-by-step process to answer a question in science.\n",
-      "a. true\n",
-      "b. false\n",
-      "b. false\n",
-      "NDQ_000023\n",
-      "series of logical steps that scientists may use to seek answers to questions\n",
-      "a. control\n",
-      "b. dependent variable\n",
-      "c. hypothesis\n",
-      "d. independent variable\n",
-      "e. physical model\n",
-      "f. theory\n",
-      "g. scientific method\n",
-      "g. scientific method\n",
-      "NDQ_000024\n",
-      "possible answer to a question that can be tested to see whether it is false\n",
-      "a. control\n",
-      "b. dependent variable\n",
-      "c. hypothesis\n",
-      "d. independent variable\n",
-      "e. physical model\n",
-      "f. theory\n",
-      "g. scientific method\n",
-      "c. hypothesis\n",
-      "NDQ_000025\n",
-      "The dependent variable in an experiment is directly influenced by the independent variable.\n",
-      "a. true\n",
-      "b. false\n",
-      "a. true\n",
-      "NDQ_000026\n",
-      "variable that is measured in an experiment to see how it is affected by another variable\n",
-      "a. control\n",
-      "b. dependent variable\n",
-      "c. hypothesis\n",
-      "d. independent variable\n",
-      "e. physical model\n",
-      "f. theory\n",
-      "g. scientific method\n",
-      "b. dependent variable\n",
-      "NDQ_000027\n",
-      "Even if there is information we dont know, a model can be used to explain an event.\n",
-      "a. true\n",
-      "b. false\n",
-      "a. true\n",
-      "NDQ_000028\n",
-      "A theory will still remain even if conflicting data is discovered.\n",
-      "a. true\n",
-      "b. false\n",
-      "b. false\n",
-      "NDQ_000036\n",
-      "Science is a set of knowledge and also a way of knowing things.\n",
-      "a. true\n",
-      "b. false\n",
-      "a. true\n",
-      "NDQ_000037\n",
-      "A scientific hypothesis is useful only if it is proven to be true.\n",
-      "a. true\n",
-      "b. false\n",
-      "b. false\n",
-      "NDQ_000038\n",
-      "Controls are factors that could affect the outcome of an experiment.\n",
-      "a. true\n",
-      "b. false\n",
-      "a. true\n",
-      "NDQ_000039\n",
-      "A scientific investigation always begins with a theory.\n",
-      "a. true\n",
-      "b. false\n",
-      "b. false\n",
-      "NDQ_000040\n",
-      "An example of a scientific hypothesis is that angels dance on the head of a pin.\n",
-      "a. true\n",
-      "b. false\n",
-      "b. false\n",
-      "NDQ_000041\n",
-      "An example of a scientific theory is the theory of plate tectonics.\n",
-      "a. true\n",
-      "b. false\n",
-      "a. true\n",
-      "NDQ_000042\n",
-      "A globe is a physical model of planet Earth.\n",
-      "a. true\n",
-      "b. false\n",
-      "a. true\n",
-      "NDQ_000043\n",
-      "Only models that are perfect representations of reality are useful in science.\n",
-      "a. true\n",
-      "b. false\n",
-      "b. false\n",
-      "NDQ_000044\n",
-      "You should wear a hoodie to protect your hair when you work in a science lab.\n",
-      "a. true\n",
-      "b. false\n",
-      "b. false\n",
-      "NDQ_000045\n",
-      "You should take a first aid kit when you do a science investigation in the field.\n",
-      "a. true\n",
-      "b. false\n",
-      "a. true\n"
-     ]
-    }
-   ],
-   "source": [
-    "# lessons = {}\n",
-    "for idx, concept in enumerate(qa_doc):\n",
-    "    print(concept[\"lessonName\"])\n",
-    "    # for key, val in concept[\"adjunctTopics\"].items():\n",
-    "    #     if key.upper() != \"VOCABULARY\":\n",
-    "    #         print(key.upper(), val[\"content\"][\"text\"])\n",
-    "    #     else:\n",
-    "    #         print(key.upper(), val)\n",
-    "    for key, val in concept[\"topics\"].items():\n",
-    "        print(key, val[\"content\"][\"text\"])\n",
-    "    print()\n",
-    "\n",
-    "    for key, val in concept[\"questions\"][\"nonDiagramQuestions\"].items():\n",
-    "        question = val[\"beingAsked\"][\"processedText\"]\n",
-    "        correct_answer = val[\"correctAnswer\"][\"processedText\"]\n",
-    "        correct_answer = val[\"answerChoices\"][correct_answer][\"rawText\"]\n",
-    "        for _, mc in val[\"answerChoices\"].items():\n",
-    "            question += \"\\n\" + mc[\"idStructural\"] + \" \" + mc[\"processedText\"]\n",
-    "        # print(key, val[\"answerChoices\"], val[\"beingAsked\"])\n",
-    "        print(key)\n",
-    "        print(question)\n",
-    "        print(correct_answer)\n",
-    "    \n",
-    "    break\n",
-    "\n",
-    "# lessons"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 1,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "import json\n",
-    "\n",
-    "file = \"/home/paperspace/Projects/posteriors/datasets/head_qa/data/HEAD_EN/test_HEAD_EN.json\"\n",
-    "with open(file, \"r\") as f:\n",
-    "    data = json.load(f)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 4,
-   "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Cuaderno_2016_1_B\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "Cuaderno_2016_1_E\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "Cuaderno_2016_1_F\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "Cuaderno_2016_1_M\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "Cuaderno_2016_1_P\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "Cuaderno_2016_1_Q\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "Cuaderno_2017_1_B\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "Cuaderno_2017_1_E\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "Cuaderno_2017_1_F\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "Cuaderno_2017_1_M\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "Cuaderno_2017_1_P\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "Cuaderno_2017_1_Q\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n",
-      "dict_keys(['qid', 'qtext', 'ra', 'answers', 'image'])\n"
-     ]
-    }
-   ],
-   "source": [
-    "for key, val in data[\"exams\"].items():\n",
-    "    print(key)\n",
-    "    for q in val[\"data\"]:\n",
-    "        print(q.keys())\n",
-    "        question = q[\"qtext\"]\n",
-    "        for mc in q[\"answers\"]:\n",
-    "            question += \"\\n\" + f\"{str(mc['aid'])}.\" + \" \" + mc[\"atext\"]\n",
-    "\n",
-    "        for answer in q[\"answers\"]:\n",
-    "            if str(answer[\"aid\"]) == str(q[\"ra\"]):\n",
-    "                atext = answer[\"atext\"]\n",
-    "        correct_answer = f\"{str(q['ra'])}.\" + \" \" + atext"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": []
-  }
- ],
- "metadata": {
-  "kernelspec": {
-   "display_name": "posteriors",
-   "language": "python",
-   "name": "python3"
-  },
-  "language_info": {
-   "codemirror_mode": {
-    "name": "ipython",
-    "version": 3
-   },
-   "file_extension": ".py",
-   "mimetype": "text/x-python",
-   "name": "python",
-   "nbconvert_exporter": "python",
-   "pygments_lexer": "ipython3",
-   "version": "3.11.9"
-  }
- },
- "nbformat": 4,
- "nbformat_minor": 2
-}
diff --git a/examples/bayes_llama3/demo.ipynb b/examples/bayes_llama3/demo.ipynb
deleted file mode 100644
index 4be12d37..00000000
--- a/examples/bayes_llama3/demo.ipynb
+++ /dev/null
@@ -1,206 +0,0 @@
-{
- "cells": [
-  {
-   "cell_type": "code",
-   "execution_count": 1,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "import torch\n",
-    "import re\n",
-    "import os\n",
-    "\n",
-    "def load_ensemble(filepaths):\n",
-    "    def load_from_checkpoint(idx, filepath):\n",
-    "        parameters = torch.load(filepath)[\"state_dict\"][\"bayesian_layer\"].params\n",
-    "        parameters = {\n",
-    "            re.sub(r\"model\\.layers\\.\\d+\\.\", \"\", k): v\n",
-    "            for k, v in parameters.items()\n",
-    "            if v.numel() > 0\n",
-    "        }\n",
-    "        return parameters\n",
-    "\n",
-    "    return [\n",
-    "        load_from_checkpoint(idx, filepath) for idx, filepath in enumerate(filepaths)\n",
-    "    ]\n",
-    "\n",
-    "\n",
-    "folder = \"logs/bayes_instruct_with_seq/usable_checkpoints\"\n",
-    "parameters = load_ensemble(\n",
-    "    [os.path.join(folder, ckpt) for ckpt in os.listdir(folder)]\n",
-    ")"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 2,
-   "metadata": {},
-   "outputs": [
-    {
-     "name": "stderr",
-     "output_type": "stream",
-     "text": [
-      "/home/paperspace/miniconda/envs/posteriors/lib/python3.11/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
-      "  from .autonotebook import tqdm as notebook_tqdm\n",
-      "Loading checkpoint shards: 100%|██████████| 4/4 [00:04<00:00,  1.04s/it]\n",
-      "Some weights of BayesLlamaForCausalLM were not initialized from the model checkpoint at Meta-Llama-3-8B-Instruct and are newly initialized: ['model.bayesian_layers.0.input_layernorm.weight', 'model.bayesian_layers.0.mlp.down_proj.weight', 'model.bayesian_layers.0.mlp.gate_proj.weight', 'model.bayesian_layers.0.mlp.up_proj.weight', 'model.bayesian_layers.0.post_attention_layernorm.weight', 'model.bayesian_layers.0.self_attn.k_proj.weight', 'model.bayesian_layers.0.self_attn.o_proj.weight', 'model.bayesian_layers.0.self_attn.q_proj.weight', 'model.bayesian_layers.0.self_attn.v_proj.weight', 'model.bayesian_layers.1.input_layernorm.weight', 'model.bayesian_layers.1.mlp.down_proj.weight', 'model.bayesian_layers.1.mlp.gate_proj.weight', 'model.bayesian_layers.1.mlp.up_proj.weight', 'model.bayesian_layers.1.post_attention_layernorm.weight', 'model.bayesian_layers.1.self_attn.k_proj.weight', 'model.bayesian_layers.1.self_attn.o_proj.weight', 'model.bayesian_layers.1.self_attn.q_proj.weight', 'model.bayesian_layers.1.self_attn.v_proj.weight', 'model.bayesian_layers.2.input_layernorm.weight', 'model.bayesian_layers.2.mlp.down_proj.weight', 'model.bayesian_layers.2.mlp.gate_proj.weight', 'model.bayesian_layers.2.mlp.up_proj.weight', 'model.bayesian_layers.2.post_attention_layernorm.weight', 'model.bayesian_layers.2.self_attn.k_proj.weight', 'model.bayesian_layers.2.self_attn.o_proj.weight', 'model.bayesian_layers.2.self_attn.q_proj.weight', 'model.bayesian_layers.2.self_attn.v_proj.weight', 'model.bayesian_layers.3.input_layernorm.weight', 'model.bayesian_layers.3.mlp.down_proj.weight', 'model.bayesian_layers.3.mlp.gate_proj.weight', 'model.bayesian_layers.3.mlp.up_proj.weight', 'model.bayesian_layers.3.post_attention_layernorm.weight', 'model.bayesian_layers.3.self_attn.k_proj.weight', 'model.bayesian_layers.3.self_attn.o_proj.weight', 'model.bayesian_layers.3.self_attn.q_proj.weight', 'model.bayesian_layers.3.self_attn.v_proj.weight', 'model.bayesian_layers.4.input_layernorm.weight', 'model.bayesian_layers.4.mlp.down_proj.weight', 'model.bayesian_layers.4.mlp.gate_proj.weight', 'model.bayesian_layers.4.mlp.up_proj.weight', 'model.bayesian_layers.4.post_attention_layernorm.weight', 'model.bayesian_layers.4.self_attn.k_proj.weight', 'model.bayesian_layers.4.self_attn.o_proj.weight', 'model.bayesian_layers.4.self_attn.q_proj.weight', 'model.bayesian_layers.4.self_attn.v_proj.weight', 'model.bayesian_layers.5.input_layernorm.weight', 'model.bayesian_layers.5.mlp.down_proj.weight', 'model.bayesian_layers.5.mlp.gate_proj.weight', 'model.bayesian_layers.5.mlp.up_proj.weight', 'model.bayesian_layers.5.post_attention_layernorm.weight', 'model.bayesian_layers.5.self_attn.k_proj.weight', 'model.bayesian_layers.5.self_attn.o_proj.weight', 'model.bayesian_layers.5.self_attn.q_proj.weight', 'model.bayesian_layers.5.self_attn.v_proj.weight', 'model.bayesian_layers.6.input_layernorm.weight', 'model.bayesian_layers.6.mlp.down_proj.weight', 'model.bayesian_layers.6.mlp.gate_proj.weight', 'model.bayesian_layers.6.mlp.up_proj.weight', 'model.bayesian_layers.6.post_attention_layernorm.weight', 'model.bayesian_layers.6.self_attn.k_proj.weight', 'model.bayesian_layers.6.self_attn.o_proj.weight', 'model.bayesian_layers.6.self_attn.q_proj.weight', 'model.bayesian_layers.6.self_attn.v_proj.weight', 'model.bayesian_layers.7.input_layernorm.weight', 'model.bayesian_layers.7.mlp.down_proj.weight', 'model.bayesian_layers.7.mlp.gate_proj.weight', 'model.bayesian_layers.7.mlp.up_proj.weight', 'model.bayesian_layers.7.post_attention_layernorm.weight', 'model.bayesian_layers.7.self_attn.k_proj.weight', 'model.bayesian_layers.7.self_attn.o_proj.weight', 'model.bayesian_layers.7.self_attn.q_proj.weight', 'model.bayesian_layers.7.self_attn.v_proj.weight', 'model.bayesian_layers.8.input_layernorm.weight', 'model.bayesian_layers.8.mlp.down_proj.weight', 'model.bayesian_layers.8.mlp.gate_proj.weight', 'model.bayesian_layers.8.mlp.up_proj.weight', 'model.bayesian_layers.8.post_attention_layernorm.weight', 'model.bayesian_layers.8.self_attn.k_proj.weight', 'model.bayesian_layers.8.self_attn.o_proj.weight', 'model.bayesian_layers.8.self_attn.q_proj.weight', 'model.bayesian_layers.8.self_attn.v_proj.weight', 'model.bayesian_layers.9.input_layernorm.weight', 'model.bayesian_layers.9.mlp.down_proj.weight', 'model.bayesian_layers.9.mlp.gate_proj.weight', 'model.bayesian_layers.9.mlp.up_proj.weight', 'model.bayesian_layers.9.post_attention_layernorm.weight', 'model.bayesian_layers.9.self_attn.k_proj.weight', 'model.bayesian_layers.9.self_attn.o_proj.weight', 'model.bayesian_layers.9.self_attn.q_proj.weight', 'model.bayesian_layers.9.self_attn.v_proj.weight']\n",
-      "You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n",
-      "Loading ensemble weights: 100%|██████████| 10/10 [00:00<00:00, 1924.70it/s]\n"
-     ]
-    }
-   ],
-   "source": [
-    "%load_ext autoreload\n",
-    "%autoreload 2\n",
-    "\n",
-    "from llama3.modules.bayesllama import BayesLlamaForCausalLM\n",
-    "bayes_config = {'n_ensemble': 10}\n",
-    "\n",
-    "bayes = BayesLlamaForCausalLM.from_pretrained(\"Meta-Llama-3-8B-Instruct\", bayes_config=bayes_config).to(\"cuda:0\")\n",
-    "bayes.load_bayesian_layers(parameters)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 5,
-   "metadata": {},
-   "outputs": [
-    {
-     "name": "stderr",
-     "output_type": "stream",
-     "text": [
-      "Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.\n"
-     ]
-    },
-    {
-     "data": {
-      "text/plain": [
-       "torch.Size([1, 5])"
-      ]
-     },
-     "execution_count": 5,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "from transformers import AutoTokenizer\n",
-    "\n",
-    "tokenizer = AutoTokenizer.from_pretrained(\"Meta-Llama-3-8B-Instruct\")\n",
-    "tokenizer.pad_token = tokenizer.eos_token\n",
-    "tokenizer.padding_side = \"left\"\n",
-    "\n",
-    "inputs = \"This is a test\"\n",
-    "inputs = tokenizer(inputs, return_tensors=\"pt\")\n",
-    "\n",
-    "inputs.input_ids.shape"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 7,
-   "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "['\\nAnswer the following multiple choice question.\\nSteps of the scientific method include all of the following except\\na. doing background research.\\nb. constructing a hypothesis.\\nc. asking a question.\\nd. proving a theory.\\nAnswer:\\n', '\\nAnswer the following multiple choice question.\\nAll birds build nests the same way.\\na. true\\nb. false\\nAnswer:\\n', '\\nAnswer the following multiple choice question.\\nIf the results of an experiment disprove a hypothesis, then the\\na. results should not be reported.\\nb. hypothesis is just a theory.\\nc. data must contain errors.\\nd. none of the above\\nAnswer:\\n']\n"
-     ]
-    },
-    {
-     "data": {
-      "text/plain": [
-       "['d', 'b', 'c. data must contain errors.  (The']"
-      ]
-     },
-     "execution_count": 7,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "@torch.no_grad()\n",
-    "def generate(model, inputs, max_length=10, use_cache=True):\n",
-    "    seq_out = []\n",
-    "    for idx in range(max_length):\n",
-    "        outputs = model(**inputs, return_dict=False, use_cache=use_cache)\n",
-    "\n",
-    "        if \"attention_mask\" in inputs:\n",
-    "            del inputs[\"attention_mask\"]\n",
-    "\n",
-    "        next_token = outputs[0][0][:, -1].argmax(-1).unsqueeze(-1)\n",
-    "        if use_cache:\n",
-    "            inputs[\"past_key_values\"] = outputs[1]\n",
-    "            inputs[\"ensemble_past_key_values\"] = outputs[2]\n",
-    "            inputs[\"input_ids\"] = next_token\n",
-    "        else:\n",
-    "            inputs[\"input_ids\"] = torch.cat([inputs[\"input_ids\"], next_token], dim=1)\n",
-    "        seq_out.append(next_token)\n",
-    "\n",
-    "    seq_out = torch.cat(seq_out, -1)\n",
-    "    return tokenizer.batch_decode(seq_out, skip_special_tokens=True)\n",
-    "\n",
-    "# inputs = \"This is a test\"\n",
-    "inputs_a = \"\"\"\n",
-    "Answer the following multiple choice question.\n",
-    "Steps of the scientific method include all of the following except\n",
-    "a. doing background research.\n",
-    "b. constructing a hypothesis.\n",
-    "c. asking a question.\n",
-    "d. proving a theory.\n",
-    "Answer:\n",
-    "\"\"\"\n",
-    "inputs_b = \"\"\"\n",
-    "Answer the following multiple choice question.\n",
-    "All birds build nests the same way.\n",
-    "a. true\n",
-    "b. false\n",
-    "Answer:\n",
-    "\"\"\"\n",
-    "inputs_c = \"\"\"\n",
-    "Answer the following multiple choice question.\n",
-    "If the results of an experiment disprove a hypothesis, then the\n",
-    "a. results should not be reported.\n",
-    "b. hypothesis is just a theory.\n",
-    "c. data must contain errors.\n",
-    "d. none of the above\n",
-    "Answer:\n",
-    "\"\"\"\n",
-    "\n",
-    "inputs = [inputs_a, inputs_b, inputs_c]\n",
-    "print(inputs)\n",
-    "inputs = tokenizer(inputs, padding=True, return_tensors=\"pt\")\n",
-    "generate(bayes, inputs.to(\"cuda\"), use_cache=True)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": []
-  }
- ],
- "metadata": {
-  "kernelspec": {
-   "display_name": "uq",
-   "language": "python",
-   "name": "python3"
-  },
-  "language_info": {
-   "codemirror_mode": {
-    "name": "ipython",
-    "version": 3
-   },
-   "file_extension": ".py",
-   "mimetype": "text/x-python",
-   "name": "python",
-   "nbconvert_exporter": "python",
-   "pygments_lexer": "ipython3",
-   "version": "3.11.9"
-  }
- },
- "nbformat": 4,
- "nbformat_minor": 2
-}
diff --git a/examples/bayes_llama3/run_eval.ipynb b/examples/bayes_llama3/run_eval.ipynb
deleted file mode 100644
index faffa3ea..00000000
--- a/examples/bayes_llama3/run_eval.ipynb
+++ /dev/null
@@ -1,267 +0,0 @@
-{
- "cells": [
-  {
-   "cell_type": "code",
-   "execution_count": 1,
-   "metadata": {},
-   "outputs": [
-    {
-     "name": "stderr",
-     "output_type": "stream",
-     "text": [
-      "/home/paperspace/miniconda/envs/posteriors/lib/python3.11/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
-      "  from .autonotebook import tqdm as notebook_tqdm\n",
-      "Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.\n",
-      "Loading checkpoint shards: 100%|██████████| 4/4 [00:03<00:00,  1.29it/s]\n"
-     ]
-    }
-   ],
-   "source": [
-    "%load_ext autoreload\n",
-    "%autoreload 2\n",
-    "\n",
-    "from llama3.eval import Experiment\n",
-    "from llama3.utils.load_utils import load_config, save_config, setup_log_dir\n",
-    "\n",
-    "config = load_config(\"configs/eval_pretrain.yaml\")\n",
-    "\n",
-    "experiment_log_dir = setup_log_dir(\n",
-    "    config.get(\"logs_dir\", \"experiment_logs\"), experiment_name=config.get(\"experiment_name\", None)\n",
-    ")\n",
-    "\n",
-    "config[\"experiment_config\"][\"experiment_log_dir\"] = experiment_log_dir\n",
-    "save_config(config.to_dict(), experiment_log_dir + \"/config.yaml\")\n",
-    "\n",
-    "experiment = Experiment(config[\"experiment_config\"])"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 25,
-   "metadata": {},
-   "outputs": [
-    {
-     "name": "stderr",
-     "output_type": "stream",
-     "text": [
-      "  0%|          | 0/210 [00:00<?, ?it/s]"
-     ]
-    },
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "b. constructing d. proving a theory.\n",
-      "c. Many c. Many scientists have agreed upon this explanation after repeated experiments and models have shown it\n",
-      "d. all d. all of the above\n",
-      "b. A b. A chart with nutritional information about food we eat.\n",
-      "d. none d. none of the above\n",
-      "a. Bring d. More than one answer is correct.\n",
-      "a. The a. The amount of damage each building receives.\n",
-      "b. A a. A theory can never be disproven.\n",
-      "d. all d. all of the above\n",
-      "d. All d. All of the above.\n",
-      "b. You b. You should tie back your hair if it is long.\n",
-      "b. dependent e. physical model\n",
-      "a. control a. control\n",
-      "b\n",
-      "Explanation d. independent variable\n",
-      "b. false b. false\n",
-      "f. theory f. theory\n",
-      "a. true b. false\n",
-      "a. control g. scientific method\n",
-      "b\n",
-      "Explanation c. hypothesis\n"
-     ]
-    },
-    {
-     "name": "stderr",
-     "output_type": "stream",
-     "text": [
-      "  0%|          | 0/210 [00:03<?, ?it/s]"
-     ]
-    },
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "b. false a. true\n"
-     ]
-    },
-    {
-     "name": "stderr",
-     "output_type": "stream",
-     "text": [
-      "\n"
-     ]
-    },
-    {
-     "ename": "KeyboardInterrupt",
-     "evalue": "",
-     "output_type": "error",
-     "traceback": [
-      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
-      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
-      "Cell \u001b[0;32mIn[25], line 3\u001b[0m\n\u001b[1;32m      1\u001b[0m dset_name \u001b[38;5;241m=\u001b[39m \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mtqa\u001b[39m\u001b[38;5;124m\"\u001b[39m\n\u001b[1;32m      2\u001b[0m experiment\u001b[38;5;241m.\u001b[39mn_tokens \u001b[38;5;241m=\u001b[39m \u001b[38;5;241m3\u001b[39m\n\u001b[0;32m----> 3\u001b[0m \u001b[43mexperiment\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mrun\u001b[49m\u001b[43m(\u001b[49m\u001b[43mdset_name\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mconfig\u001b[49m\u001b[43m[\u001b[49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mdataset_path\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m]\u001b[49m\u001b[43m[\u001b[49m\u001b[43mdset_name\u001b[49m\u001b[43m]\u001b[49m\u001b[43m[\u001b[49m\u001b[38;5;241;43m0\u001b[39;49m\u001b[43m]\u001b[49m\u001b[43m)\u001b[49m\n",
-      "File \u001b[0;32m~/Projects/posteriors/examples/bayes_llama3/llama3/eval.py:244\u001b[0m, in \u001b[0;36mExperiment.run\u001b[0;34m(self, dataset_name, dataset_path, **kwargs)\u001b[0m\n\u001b[1;32m    242\u001b[0m \u001b[38;5;28;01melif\u001b[39;00m \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mtest\u001b[39m\u001b[38;5;124m\"\u001b[39m \u001b[38;5;129;01min\u001b[39;00m splits:\n\u001b[1;32m    243\u001b[0m     split \u001b[38;5;241m=\u001b[39m \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mtest\u001b[39m\u001b[38;5;124m\"\u001b[39m\n\u001b[0;32m--> 244\u001b[0m results \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mrun_experiment\u001b[49m\u001b[43m(\u001b[49m\u001b[43mdataset_name\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mdataset_path\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43msplit\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43mkwargs\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    245\u001b[0m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39msave_results(dataset_name, results, split)\n\u001b[1;32m    247\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m results\n",
-      "File \u001b[0;32m~/Projects/posteriors/examples/bayes_llama3/llama3/eval.py:191\u001b[0m, in \u001b[0;36mExperiment.run_experiment\u001b[0;34m(self, dataset_name, dataset_path, split)\u001b[0m\n\u001b[1;32m    187\u001b[0m inputs \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mprepare_prompt(PROMPT, questions)\n\u001b[1;32m    189\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39meval_pretrained:\n\u001b[1;32m    190\u001b[0m     answers, total_uncertainties, epistemic_uncertainties \u001b[38;5;241m=\u001b[39m (\n\u001b[0;32m--> 191\u001b[0m         \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mgenerate_base\u001b[49m\u001b[43m(\u001b[49m\n\u001b[1;32m    192\u001b[0m \u001b[43m            \u001b[49m\u001b[43minputs\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mto\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mmodel\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mdevice\u001b[49m\u001b[43m)\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mmax_length\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mn_tokens\u001b[49m\n\u001b[1;32m    193\u001b[0m \u001b[43m        \u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    194\u001b[0m     )\n\u001b[1;32m    195\u001b[0m \u001b[38;5;28;01melse\u001b[39;00m:\n\u001b[1;32m    196\u001b[0m     answers, total_uncertainties, epistemic_uncertainties \u001b[38;5;241m=\u001b[39m (\n\u001b[1;32m    197\u001b[0m         \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mgenerate(\n\u001b[1;32m    198\u001b[0m             inputs\u001b[38;5;241m.\u001b[39mto(\u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mmodel\u001b[38;5;241m.\u001b[39mdevice), max_length\u001b[38;5;241m=\u001b[39m\u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mn_tokens\n\u001b[1;32m    199\u001b[0m         )\n\u001b[1;32m    200\u001b[0m     )\n",
-      "File \u001b[0;32m~/miniconda/envs/posteriors/lib/python3.11/site-packages/torch/utils/_contextlib.py:115\u001b[0m, in \u001b[0;36mcontext_decorator.<locals>.decorate_context\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m    112\u001b[0m \u001b[38;5;129m@functools\u001b[39m\u001b[38;5;241m.\u001b[39mwraps(func)\n\u001b[1;32m    113\u001b[0m \u001b[38;5;28;01mdef\u001b[39;00m \u001b[38;5;21mdecorate_context\u001b[39m(\u001b[38;5;241m*\u001b[39margs, \u001b[38;5;241m*\u001b[39m\u001b[38;5;241m*\u001b[39mkwargs):\n\u001b[1;32m    114\u001b[0m     \u001b[38;5;28;01mwith\u001b[39;00m ctx_factory():\n\u001b[0;32m--> 115\u001b[0m         \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[43mfunc\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43margs\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43mkwargs\u001b[49m\u001b[43m)\u001b[49m\n",
-      "File \u001b[0;32m~/Projects/posteriors/examples/bayes_llama3/llama3/eval.py:140\u001b[0m, in \u001b[0;36mExperiment.generate_base\u001b[0;34m(self, inputs, max_length, use_cache)\u001b[0m\n\u001b[1;32m    138\u001b[0m total_uncertainties \u001b[38;5;241m=\u001b[39m [[] \u001b[38;5;28;01mfor\u001b[39;00m _ \u001b[38;5;129;01min\u001b[39;00m \u001b[38;5;28mrange\u001b[39m(inputs[\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124minput_ids\u001b[39m\u001b[38;5;124m\"\u001b[39m]\u001b[38;5;241m.\u001b[39msize(\u001b[38;5;241m0\u001b[39m))]\n\u001b[1;32m    139\u001b[0m \u001b[38;5;28;01mfor\u001b[39;00m _ \u001b[38;5;129;01min\u001b[39;00m \u001b[38;5;28mrange\u001b[39m(max_length):\n\u001b[0;32m--> 140\u001b[0m     outputs \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mmodel\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43minputs\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mreturn_dict\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;28;43;01mFalse\u001b[39;49;00m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43muse_cache\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43muse_cache\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    142\u001b[0m     \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mattention_mask\u001b[39m\u001b[38;5;124m\"\u001b[39m \u001b[38;5;129;01min\u001b[39;00m inputs:\n\u001b[1;32m    143\u001b[0m         \u001b[38;5;28;01mdel\u001b[39;00m inputs[\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mattention_mask\u001b[39m\u001b[38;5;124m\"\u001b[39m]\n",
-      "File \u001b[0;32m~/miniconda/envs/posteriors/lib/python3.11/site-packages/torch/nn/modules/module.py:1511\u001b[0m, in \u001b[0;36mModule._wrapped_call_impl\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m   1509\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_compiled_call_impl(\u001b[38;5;241m*\u001b[39margs, \u001b[38;5;241m*\u001b[39m\u001b[38;5;241m*\u001b[39mkwargs)  \u001b[38;5;66;03m# type: ignore[misc]\u001b[39;00m\n\u001b[1;32m   1510\u001b[0m \u001b[38;5;28;01melse\u001b[39;00m:\n\u001b[0;32m-> 1511\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_call_impl\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43margs\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43mkwargs\u001b[49m\u001b[43m)\u001b[49m\n",
-      "File \u001b[0;32m~/miniconda/envs/posteriors/lib/python3.11/site-packages/torch/nn/modules/module.py:1520\u001b[0m, in \u001b[0;36mModule._call_impl\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m   1515\u001b[0m \u001b[38;5;66;03m# If we don't have any hooks, we want to skip the rest of the logic in\u001b[39;00m\n\u001b[1;32m   1516\u001b[0m \u001b[38;5;66;03m# this function, and just call forward.\u001b[39;00m\n\u001b[1;32m   1517\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m (\u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_backward_hooks \u001b[38;5;129;01mor\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_backward_pre_hooks \u001b[38;5;129;01mor\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_forward_hooks \u001b[38;5;129;01mor\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_forward_pre_hooks\n\u001b[1;32m   1518\u001b[0m         \u001b[38;5;129;01mor\u001b[39;00m _global_backward_pre_hooks \u001b[38;5;129;01mor\u001b[39;00m _global_backward_hooks\n\u001b[1;32m   1519\u001b[0m         \u001b[38;5;129;01mor\u001b[39;00m _global_forward_hooks \u001b[38;5;129;01mor\u001b[39;00m _global_forward_pre_hooks):\n\u001b[0;32m-> 1520\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[43mforward_call\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43margs\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43mkwargs\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m   1522\u001b[0m \u001b[38;5;28;01mtry\u001b[39;00m:\n\u001b[1;32m   1523\u001b[0m     result \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;01mNone\u001b[39;00m\n",
-      "File \u001b[0;32m~/miniconda/envs/posteriors/lib/python3.11/site-packages/transformers/models/llama/modeling_llama.py:1208\u001b[0m, in \u001b[0;36mLlamaForCausalLM.forward\u001b[0;34m(self, input_ids, attention_mask, position_ids, past_key_values, inputs_embeds, labels, use_cache, output_attentions, output_hidden_states, return_dict, cache_position)\u001b[0m\n\u001b[1;32m   1205\u001b[0m return_dict \u001b[38;5;241m=\u001b[39m return_dict \u001b[38;5;28;01mif\u001b[39;00m return_dict \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m \u001b[38;5;28;01melse\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mconfig\u001b[38;5;241m.\u001b[39muse_return_dict\n\u001b[1;32m   1207\u001b[0m \u001b[38;5;66;03m# decoder outputs consists of (dec_features, layer_state, dec_hidden, dec_attn)\u001b[39;00m\n\u001b[0;32m-> 1208\u001b[0m outputs \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mmodel\u001b[49m\u001b[43m(\u001b[49m\n\u001b[1;32m   1209\u001b[0m \u001b[43m    \u001b[49m\u001b[43minput_ids\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43minput_ids\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m   1210\u001b[0m \u001b[43m    \u001b[49m\u001b[43mattention_mask\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mattention_mask\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m   1211\u001b[0m \u001b[43m    \u001b[49m\u001b[43mposition_ids\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mposition_ids\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m   1212\u001b[0m \u001b[43m    \u001b[49m\u001b[43mpast_key_values\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mpast_key_values\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m   1213\u001b[0m \u001b[43m    \u001b[49m\u001b[43minputs_embeds\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43minputs_embeds\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m   1214\u001b[0m \u001b[43m    \u001b[49m\u001b[43muse_cache\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43muse_cache\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m   1215\u001b[0m \u001b[43m    \u001b[49m\u001b[43moutput_attentions\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43moutput_attentions\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m   1216\u001b[0m \u001b[43m    \u001b[49m\u001b[43moutput_hidden_states\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43moutput_hidden_states\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m   1217\u001b[0m \u001b[43m    \u001b[49m\u001b[43mreturn_dict\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mreturn_dict\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m   1218\u001b[0m \u001b[43m    \u001b[49m\u001b[43mcache_position\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mcache_position\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m   1219\u001b[0m \u001b[43m\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m   1221\u001b[0m hidden_states \u001b[38;5;241m=\u001b[39m outputs[\u001b[38;5;241m0\u001b[39m]\n\u001b[1;32m   1222\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mconfig\u001b[38;5;241m.\u001b[39mpretraining_tp \u001b[38;5;241m>\u001b[39m \u001b[38;5;241m1\u001b[39m:\n",
-      "File \u001b[0;32m~/miniconda/envs/posteriors/lib/python3.11/site-packages/torch/nn/modules/module.py:1511\u001b[0m, in \u001b[0;36mModule._wrapped_call_impl\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m   1509\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_compiled_call_impl(\u001b[38;5;241m*\u001b[39margs, \u001b[38;5;241m*\u001b[39m\u001b[38;5;241m*\u001b[39mkwargs)  \u001b[38;5;66;03m# type: ignore[misc]\u001b[39;00m\n\u001b[1;32m   1510\u001b[0m \u001b[38;5;28;01melse\u001b[39;00m:\n\u001b[0;32m-> 1511\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_call_impl\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43margs\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43mkwargs\u001b[49m\u001b[43m)\u001b[49m\n",
-      "File \u001b[0;32m~/miniconda/envs/posteriors/lib/python3.11/site-packages/torch/nn/modules/module.py:1520\u001b[0m, in \u001b[0;36mModule._call_impl\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m   1515\u001b[0m \u001b[38;5;66;03m# If we don't have any hooks, we want to skip the rest of the logic in\u001b[39;00m\n\u001b[1;32m   1516\u001b[0m \u001b[38;5;66;03m# this function, and just call forward.\u001b[39;00m\n\u001b[1;32m   1517\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m (\u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_backward_hooks \u001b[38;5;129;01mor\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_backward_pre_hooks \u001b[38;5;129;01mor\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_forward_hooks \u001b[38;5;129;01mor\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_forward_pre_hooks\n\u001b[1;32m   1518\u001b[0m         \u001b[38;5;129;01mor\u001b[39;00m _global_backward_pre_hooks \u001b[38;5;129;01mor\u001b[39;00m _global_backward_hooks\n\u001b[1;32m   1519\u001b[0m         \u001b[38;5;129;01mor\u001b[39;00m _global_forward_hooks \u001b[38;5;129;01mor\u001b[39;00m _global_forward_pre_hooks):\n\u001b[0;32m-> 1520\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[43mforward_call\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43margs\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43mkwargs\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m   1522\u001b[0m \u001b[38;5;28;01mtry\u001b[39;00m:\n\u001b[1;32m   1523\u001b[0m     result \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;01mNone\u001b[39;00m\n",
-      "File \u001b[0;32m~/miniconda/envs/posteriors/lib/python3.11/site-packages/transformers/models/llama/modeling_llama.py:1018\u001b[0m, in \u001b[0;36mLlamaModel.forward\u001b[0;34m(self, input_ids, attention_mask, position_ids, past_key_values, inputs_embeds, use_cache, output_attentions, output_hidden_states, return_dict, cache_position)\u001b[0m\n\u001b[1;32m   1007\u001b[0m     layer_outputs \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_gradient_checkpointing_func(\n\u001b[1;32m   1008\u001b[0m         decoder_layer\u001b[38;5;241m.\u001b[39m\u001b[38;5;21m__call__\u001b[39m,\n\u001b[1;32m   1009\u001b[0m         hidden_states,\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m   1015\u001b[0m         cache_position,\n\u001b[1;32m   1016\u001b[0m     )\n\u001b[1;32m   1017\u001b[0m \u001b[38;5;28;01melse\u001b[39;00m:\n\u001b[0;32m-> 1018\u001b[0m     layer_outputs \u001b[38;5;241m=\u001b[39m \u001b[43mdecoder_layer\u001b[49m\u001b[43m(\u001b[49m\n\u001b[1;32m   1019\u001b[0m \u001b[43m        \u001b[49m\u001b[43mhidden_states\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m   1020\u001b[0m \u001b[43m        \u001b[49m\u001b[43mattention_mask\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mcausal_mask\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m   1021\u001b[0m \u001b[43m        \u001b[49m\u001b[43mposition_ids\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mposition_ids\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m   1022\u001b[0m \u001b[43m        \u001b[49m\u001b[43mpast_key_value\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mpast_key_values\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m   1023\u001b[0m \u001b[43m        \u001b[49m\u001b[43moutput_attentions\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43moutput_attentions\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m   1024\u001b[0m \u001b[43m        \u001b[49m\u001b[43muse_cache\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43muse_cache\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m   1025\u001b[0m \u001b[43m        \u001b[49m\u001b[43mcache_position\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mcache_position\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m   1026\u001b[0m \u001b[43m    \u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m   1028\u001b[0m hidden_states \u001b[38;5;241m=\u001b[39m layer_outputs[\u001b[38;5;241m0\u001b[39m]\n\u001b[1;32m   1030\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m use_cache:\n",
-      "File \u001b[0;32m~/miniconda/envs/posteriors/lib/python3.11/site-packages/torch/nn/modules/module.py:1511\u001b[0m, in \u001b[0;36mModule._wrapped_call_impl\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m   1509\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_compiled_call_impl(\u001b[38;5;241m*\u001b[39margs, \u001b[38;5;241m*\u001b[39m\u001b[38;5;241m*\u001b[39mkwargs)  \u001b[38;5;66;03m# type: ignore[misc]\u001b[39;00m\n\u001b[1;32m   1510\u001b[0m \u001b[38;5;28;01melse\u001b[39;00m:\n\u001b[0;32m-> 1511\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_call_impl\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43margs\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43mkwargs\u001b[49m\u001b[43m)\u001b[49m\n",
-      "File \u001b[0;32m~/miniconda/envs/posteriors/lib/python3.11/site-packages/torch/nn/modules/module.py:1520\u001b[0m, in \u001b[0;36mModule._call_impl\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m   1515\u001b[0m \u001b[38;5;66;03m# If we don't have any hooks, we want to skip the rest of the logic in\u001b[39;00m\n\u001b[1;32m   1516\u001b[0m \u001b[38;5;66;03m# this function, and just call forward.\u001b[39;00m\n\u001b[1;32m   1517\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m (\u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_backward_hooks \u001b[38;5;129;01mor\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_backward_pre_hooks \u001b[38;5;129;01mor\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_forward_hooks \u001b[38;5;129;01mor\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_forward_pre_hooks\n\u001b[1;32m   1518\u001b[0m         \u001b[38;5;129;01mor\u001b[39;00m _global_backward_pre_hooks \u001b[38;5;129;01mor\u001b[39;00m _global_backward_hooks\n\u001b[1;32m   1519\u001b[0m         \u001b[38;5;129;01mor\u001b[39;00m _global_forward_hooks \u001b[38;5;129;01mor\u001b[39;00m _global_forward_pre_hooks):\n\u001b[0;32m-> 1520\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[43mforward_call\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43margs\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43mkwargs\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m   1522\u001b[0m \u001b[38;5;28;01mtry\u001b[39;00m:\n\u001b[1;32m   1523\u001b[0m     result \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;01mNone\u001b[39;00m\n",
-      "File \u001b[0;32m~/miniconda/envs/posteriors/lib/python3.11/site-packages/transformers/models/llama/modeling_llama.py:756\u001b[0m, in \u001b[0;36mLlamaDecoderLayer.forward\u001b[0;34m(self, hidden_states, attention_mask, position_ids, past_key_value, output_attentions, use_cache, cache_position, **kwargs)\u001b[0m\n\u001b[1;32m    754\u001b[0m residual \u001b[38;5;241m=\u001b[39m hidden_states\n\u001b[1;32m    755\u001b[0m hidden_states \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mpost_attention_layernorm(hidden_states)\n\u001b[0;32m--> 756\u001b[0m hidden_states \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mmlp\u001b[49m\u001b[43m(\u001b[49m\u001b[43mhidden_states\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    757\u001b[0m hidden_states \u001b[38;5;241m=\u001b[39m residual \u001b[38;5;241m+\u001b[39m hidden_states\n\u001b[1;32m    759\u001b[0m outputs \u001b[38;5;241m=\u001b[39m (hidden_states,)\n",
-      "File \u001b[0;32m~/miniconda/envs/posteriors/lib/python3.11/site-packages/torch/nn/modules/module.py:1511\u001b[0m, in \u001b[0;36mModule._wrapped_call_impl\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m   1509\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_compiled_call_impl(\u001b[38;5;241m*\u001b[39margs, \u001b[38;5;241m*\u001b[39m\u001b[38;5;241m*\u001b[39mkwargs)  \u001b[38;5;66;03m# type: ignore[misc]\u001b[39;00m\n\u001b[1;32m   1510\u001b[0m \u001b[38;5;28;01melse\u001b[39;00m:\n\u001b[0;32m-> 1511\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_call_impl\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43margs\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43mkwargs\u001b[49m\u001b[43m)\u001b[49m\n",
-      "File \u001b[0;32m~/miniconda/envs/posteriors/lib/python3.11/site-packages/torch/nn/modules/module.py:1520\u001b[0m, in \u001b[0;36mModule._call_impl\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m   1515\u001b[0m \u001b[38;5;66;03m# If we don't have any hooks, we want to skip the rest of the logic in\u001b[39;00m\n\u001b[1;32m   1516\u001b[0m \u001b[38;5;66;03m# this function, and just call forward.\u001b[39;00m\n\u001b[1;32m   1517\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m (\u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_backward_hooks \u001b[38;5;129;01mor\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_backward_pre_hooks \u001b[38;5;129;01mor\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_forward_hooks \u001b[38;5;129;01mor\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_forward_pre_hooks\n\u001b[1;32m   1518\u001b[0m         \u001b[38;5;129;01mor\u001b[39;00m _global_backward_pre_hooks \u001b[38;5;129;01mor\u001b[39;00m _global_backward_hooks\n\u001b[1;32m   1519\u001b[0m         \u001b[38;5;129;01mor\u001b[39;00m _global_forward_hooks \u001b[38;5;129;01mor\u001b[39;00m _global_forward_pre_hooks):\n\u001b[0;32m-> 1520\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[43mforward_call\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43margs\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43mkwargs\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m   1522\u001b[0m \u001b[38;5;28;01mtry\u001b[39;00m:\n\u001b[1;32m   1523\u001b[0m     result \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;01mNone\u001b[39;00m\n",
-      "File \u001b[0;32m~/miniconda/envs/posteriors/lib/python3.11/site-packages/transformers/models/llama/modeling_llama.py:240\u001b[0m, in \u001b[0;36mLlamaMLP.forward\u001b[0;34m(self, x)\u001b[0m\n\u001b[1;32m    238\u001b[0m     down_proj \u001b[38;5;241m=\u001b[39m \u001b[38;5;28msum\u001b[39m(down_proj)\n\u001b[1;32m    239\u001b[0m \u001b[38;5;28;01melse\u001b[39;00m:\n\u001b[0;32m--> 240\u001b[0m     down_proj \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mdown_proj\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mact_fn\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mgate_proj\u001b[49m\u001b[43m(\u001b[49m\u001b[43mx\u001b[49m\u001b[43m)\u001b[49m\u001b[43m)\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43m \u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mup_proj\u001b[49m\u001b[43m(\u001b[49m\u001b[43mx\u001b[49m\u001b[43m)\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    242\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m down_proj\n",
-      "File \u001b[0;32m~/miniconda/envs/posteriors/lib/python3.11/site-packages/torch/nn/modules/module.py:1511\u001b[0m, in \u001b[0;36mModule._wrapped_call_impl\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m   1509\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_compiled_call_impl(\u001b[38;5;241m*\u001b[39margs, \u001b[38;5;241m*\u001b[39m\u001b[38;5;241m*\u001b[39mkwargs)  \u001b[38;5;66;03m# type: ignore[misc]\u001b[39;00m\n\u001b[1;32m   1510\u001b[0m \u001b[38;5;28;01melse\u001b[39;00m:\n\u001b[0;32m-> 1511\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_call_impl\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43margs\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43mkwargs\u001b[49m\u001b[43m)\u001b[49m\n",
-      "File \u001b[0;32m~/miniconda/envs/posteriors/lib/python3.11/site-packages/torch/nn/modules/module.py:1520\u001b[0m, in \u001b[0;36mModule._call_impl\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m   1515\u001b[0m \u001b[38;5;66;03m# If we don't have any hooks, we want to skip the rest of the logic in\u001b[39;00m\n\u001b[1;32m   1516\u001b[0m \u001b[38;5;66;03m# this function, and just call forward.\u001b[39;00m\n\u001b[1;32m   1517\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m (\u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_backward_hooks \u001b[38;5;129;01mor\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_backward_pre_hooks \u001b[38;5;129;01mor\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_forward_hooks \u001b[38;5;129;01mor\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_forward_pre_hooks\n\u001b[1;32m   1518\u001b[0m         \u001b[38;5;129;01mor\u001b[39;00m _global_backward_pre_hooks \u001b[38;5;129;01mor\u001b[39;00m _global_backward_hooks\n\u001b[1;32m   1519\u001b[0m         \u001b[38;5;129;01mor\u001b[39;00m _global_forward_hooks \u001b[38;5;129;01mor\u001b[39;00m _global_forward_pre_hooks):\n\u001b[0;32m-> 1520\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[43mforward_call\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43margs\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43mkwargs\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m   1522\u001b[0m \u001b[38;5;28;01mtry\u001b[39;00m:\n\u001b[1;32m   1523\u001b[0m     result \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;01mNone\u001b[39;00m\n",
-      "File \u001b[0;32m~/miniconda/envs/posteriors/lib/python3.11/site-packages/torch/nn/modules/linear.py:116\u001b[0m, in \u001b[0;36mLinear.forward\u001b[0;34m(self, input)\u001b[0m\n\u001b[1;32m    115\u001b[0m \u001b[38;5;28;01mdef\u001b[39;00m \u001b[38;5;21mforward\u001b[39m(\u001b[38;5;28mself\u001b[39m, \u001b[38;5;28minput\u001b[39m: Tensor) \u001b[38;5;241m-\u001b[39m\u001b[38;5;241m>\u001b[39m Tensor:\n\u001b[0;32m--> 116\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[43mF\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mlinear\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;28;43minput\u001b[39;49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mweight\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mbias\u001b[49m\u001b[43m)\u001b[49m\n",
-      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
-     ]
-    }
-   ],
-   "source": [
-    "dset_name = \"tqa\"\n",
-    "experiment.n_tokens = 3\n",
-    "experiment.run(dset_name, config[\"dataset_path\"][dset_name][0])"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 20,
-   "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "['\\nAnswer the following multiple choice question as succintly as possible.\\n\\nSteps of the scientific method include all of the following except\\na. doing background research.\\nb. constructing a hypothesis.\\nc. asking a question.\\nd. proving a theory.\\n\\nAnswer:\\n', '\\nAnswer the following multiple choice question as succintly as possible.\\n\\nWhy do scientists call the Big Bang a theory?\\na. It is probably unlikely and therefore not a fact.\\nb. A very well respected scientist proved it to be true.\\nc. Many scientists have agreed upon this explanation after repeated experiments and models have shown it\\nd. All possible answers to a scientific idea are called theories.\\n\\nAnswer:\\n', '\\nAnswer the following multiple choice question as succintly as possible.\\n\\nThe data collected in an experiment should always be\\na. labeled.\\nb. recorded.\\nc. reported.\\nd. all of the above\\n\\nAnswer:\\n', '\\nAnswer the following multiple choice question as succintly as possible.\\n\\nIf the results of an experiment disprove a hypothesis, then the\\na. results should not be reported.\\nb. hypothesis is just a theory.\\nc. data must contain errors.\\nd. none of the above\\n\\nAnswer:\\n']\n"
-     ]
-    },
-    {
-     "data": {
-      "text/plain": [
-       "['b. constructing a hypothesis.\\n\\nExplanation:\\nThe scientific method is a process for experimentation that is used to',\n",
-       " 'd. All possible answers to a scientific idea are called theories.\\nExplanation:\\nThe Big Bang theory is',\n",
-       " 'd. all of the above\\nExplanation:\\nThe data collected in an experiment should always be labeled,',\n",
-       " 'd. none of the above\\nExplanation:\\nThe results of an experiment disprove a hypothesis, then']"
-      ]
-     },
-     "execution_count": 20,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
-   "source": [
-    "import torch\n",
-    "\n",
-    "inputs_a = \"\"\"\n",
-    "Answer the following multiple choice question as succintly as possible.\n",
-    "\n",
-    "Steps of the scientific method include all of the following except\n",
-    "a. doing background research.\n",
-    "b. constructing a hypothesis.\n",
-    "c. asking a question.\n",
-    "d. proving a theory.\n",
-    "\n",
-    "Answer:\n",
-    "\"\"\"\n",
-    "inputs_b = \"\"\"\n",
-    "Answer the following multiple choice question as succintly as possible.\n",
-    "\n",
-    "Why do scientists call the Big Bang a theory?\n",
-    "a. It is probably unlikely and therefore not a fact.\n",
-    "b. A very well respected scientist proved it to be true.\n",
-    "c. Many scientists have agreed upon this explanation after repeated experiments and models have shown it\n",
-    "d. All possible answers to a scientific idea are called theories.\n",
-    "\n",
-    "Answer:\n",
-    "\"\"\"\n",
-    "inputs_c = \"\"\"\n",
-    "Answer the following multiple choice question as succintly as possible.\n",
-    "\n",
-    "The data collected in an experiment should always be\n",
-    "a. labeled.\n",
-    "b. recorded.\n",
-    "c. reported.\n",
-    "d. all of the above\n",
-    "\n",
-    "Answer:\n",
-    "\"\"\"\n",
-    "inputs_d = \"\"\"\n",
-    "Answer the following multiple choice question as succintly as possible.\n",
-    "\n",
-    "If the results of an experiment disprove a hypothesis, then the\n",
-    "a. results should not be reported.\n",
-    "b. hypothesis is just a theory.\n",
-    "c. data must contain errors.\n",
-    "d. none of the above\n",
-    "\n",
-    "Answer:\n",
-    "\"\"\"\n",
-    "\n",
-    "@torch.no_grad()\n",
-    "def generate(model, inputs, max_length=20, is_ensemble_model=False, use_cache=True):\n",
-    "    seq_out = []\n",
-    "    for idx in range(max_length):\n",
-    "        outputs = model(**inputs, return_dict=False, use_cache=use_cache)\n",
-    "\n",
-    "        if \"attention_mask\" in inputs:\n",
-    "            del inputs[\"attention_mask\"]\n",
-    "\n",
-    "        next_token = outputs[0][:, -1].argmax(-1).unsqueeze(-1)\n",
-    "\n",
-    "        if use_cache:\n",
-    "            inputs[\"past_key_values\"] = outputs[1]\n",
-    "            if is_ensemble_model: \n",
-    "                inputs[\"ensemble_past_key_values\"] = outputs[2]\n",
-    "            inputs[\"input_ids\"] = next_token\n",
-    "        else:\n",
-    "            inputs[\"input_ids\"] = torch.cat([inputs[\"input_ids\"], next_token], dim=1)\n",
-    "        seq_out.append(next_token)\n",
-    "\n",
-    "    return torch.cat(seq_out, -1)\n",
-    "\n",
-    "inputs = [inputs_a, inputs_b, inputs_c, inputs_d]\n",
-    "print(inputs)\n",
-    "inputs = experiment.tokenizer(inputs, padding=True, return_tensors=\"pt\")\n",
-    "seq_out = generate(experiment.model, inputs.to(\"cuda\"), is_ensemble_model=not experiment.eval_pretrained, use_cache=True)\n",
-    "experiment.tokenizer.batch_decode(seq_out, skip_special_tokens=True)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": []
-  }
- ],
- "metadata": {
-  "kernelspec": {
-   "display_name": "posteriors",
-   "language": "python",
-   "name": "python3"
-  },
-  "language_info": {
-   "codemirror_mode": {
-    "name": "ipython",
-    "version": 3
-   },
-   "file_extension": ".py",
-   "mimetype": "text/x-python",
-   "name": "python",
-   "nbconvert_exporter": "python",
-   "pygments_lexer": "ipython3",
-   "version": "3.11.9"
-  }
- },
- "nbformat": 4,
- "nbformat_minor": 2
-}