create repo

.gitignore

@@ -0,0 +1,5 @@
+# Ignore local folders
+data
+env
+__pycache__
+.DS_Store

README.md

@@ -0,0 +1,6 @@
+# Maze Extrapolation
+
+## Models
+The current models are:
+- `dt_net`: a model from [Bansal et. al. ](https://github.com/aks2203/deep-thinking)
+

notebooks/check_neighbor_counting.ipynb

@@ -0,0 +1,81 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Import necessary modules\n",
+    "\n",
+    "import sys\n",
+    "import os\n",
+    "\n",
+    "# Set root folder to project root\n",
+    "os.chdir(os.path.dirname(os.getcwd()))\n",
+    "\n",
+    "# Add root folder to path\n",
+    "sys.path.append(os.getcwd())\n",
+    "\n",
+    "from src.utils.loading import get_mazes\n",
+    "from src.utils.plotting import plot_mazes"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "1.0\n"
+     ]
+    }
+   ],
+   "source": [
+    "# Load mazes\n",
+    "\n",
+    "from src.utils.testing import count_start_neighbors\n",
+    "\n",
+    "inputs, _ = get_mazes(dataset='maze-dataset', maze_size=9, num_mazes=100, deadend_start=False)\n",
+    "\n",
+    "neighbors = count_start_neighbors(inputs)\n",
+    "\n",
+    "maze_idx_start = 0\n",
+    "maze_idx_end = 10\n",
+    "print(neighbors[maze_idx_start:maze_idx_end])\n",
+    "plot_mazes(inputs[maze_idx_start:maze_idx_end])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "env",
+   "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": "3.10.14"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

notebooks/explore_failed_predictions.ipynb

@@ -0,0 +1,143 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Import necessary modules\n",
+    "\n",
+    "import sys\n",
+    "import os\n",
+    "\n",
+    "# Set root folder to project root\n",
+    "os.chdir(os.path.dirname(os.getcwd()))\n",
+    "\n",
+    "# Add root folder to path\n",
+    "sys.path.append(os.getcwd())\n",
+    "\n",
+    "import numpy as np\n",
+    "import torch\n",
+    "from matplotlib import pyplot as plt\n",
+    "\n",
+    "from src.utils.testing import compare_mazes\n",
+    "from src.utils.loading import load_model, get_mazes\n",
+    "from src.utils.plotting import plot_mazes"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Using device: cuda\n",
+      "Loaded pi_net to cuda\n"
+     ]
+    }
+   ],
+   "source": [
+    "# Load model and mazes\n",
+    "\n",
+    "model = load_model('dt_net')\n",
+    "\n",
+    "inputs, solutions = get_mazes(\n",
+    "    dataset='maze-dataset', \n",
+    "    maze_size=9, \n",
+    "    num_mazes=30,\n",
+    "    percolation=0.0,\n",
+    "    deadend_start=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Compute predictions and compare to solutions\n",
+    "\n",
+    "predictions = torch.zeros_like(solutions)\n",
+    "for i in range(inputs.shape[0]):\n",
+    "    predictions[i:i+1] = model.predict(inputs[i:i+1], iters=300)\n",
+    "\n",
+    "corrects = torch.tensor(compare_mazes(predictions, solutions), dtype=torch.bool)\n",
+    "incorrects = ~corrects"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "No incorrect predictions found.\n"
+     ]
+    }
+   ],
+   "source": [
+    "# Plot incorrect predictions\n",
+    "\n",
+    "if incorrects.any():\n",
+    "    plot_mazes(inputs[incorrects], \n",
+    "            predictions=predictions[incorrects], \n",
+    "            solutions=solutions[incorrects], \n",
+    "            file_name=f'outputs/mazes/{model.name()}_incorrects.pdf')\n",
+    "else:\n",
+    "    print('No incorrect predictions found.')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "metadata": {},
+   "outputs": [
+    {
+     "ename": "IndexError",
+     "evalue": "index 0 is out of bounds for axis 0 with size 0",
+     "output_type": "error",
+     "traceback": [
+      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
+      "\u001b[0;31mIndexError\u001b[0m                                Traceback (most recent call last)",
+      "Cell \u001b[0;32mIn[5], line 4\u001b[0m\n\u001b[1;32m      1\u001b[0m incorrect_inputs \u001b[38;5;241m=\u001b[39m inputs[incorrects]\u001b[38;5;241m.\u001b[39mcpu()\u001b[38;5;241m.\u001b[39mnumpy()\n\u001b[1;32m      2\u001b[0m incorrect_inputs \u001b[38;5;241m=\u001b[39m np\u001b[38;5;241m.\u001b[39mmoveaxis(incorrect_inputs, \u001b[38;5;241m1\u001b[39m, \u001b[38;5;241m-\u001b[39m\u001b[38;5;241m1\u001b[39m) \u001b[38;5;66;03m# Move RGB axis to last\u001b[39;00m\n\u001b[0;32m----> 4\u001b[0m plt\u001b[38;5;241m.\u001b[39mimshow(\u001b[43mincorrect_inputs\u001b[49m\u001b[43m[\u001b[49m\u001b[38;5;241;43m0\u001b[39;49m\u001b[43m]\u001b[49m, cmap\u001b[38;5;241m=\u001b[39m\u001b[38;5;124m'\u001b[39m\u001b[38;5;124mgray\u001b[39m\u001b[38;5;124m'\u001b[39m)\n",
+      "\u001b[0;31mIndexError\u001b[0m: index 0 is out of bounds for axis 0 with size 0"
+     ]
+    }
+   ],
+   "source": [
+    "incorrect_inputs = inputs[incorrects].cpu().numpy()\n",
+    "incorrect_inputs = np.moveaxis(incorrect_inputs, 1, -1) # Move RGB axis to last\n",
+    "\n",
+    "plt.imshow(incorrect_inputs[0], cmap='gray')"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "env",
+   "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": "3.10.14"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}