A01_model_openai.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "507788fa-cb65-4c26-a3b6-b933f8d82d4d",
   "metadata": {},
   "source": [
    "## Langchain LLM - OPENAI"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b706cbb5-17e6-4a3f-aac6-d3a0903acf40",
   "metadata": {},
   "source": [
    "1. OPENAI\n",
    "查看是否有額度\n",
    "- https://platform.openai.com/account/billing/overview\n",
    "- https://platform.openai.com/usage\n",
    "\n",
    "2. LANGCHAIN 手冊\n",
    "https://python.langchain.com/docs/integrations/llms/openai\n",
    "\n",
    "3. 學習llm 模組, 如 llm(\"什麼是聯邦式學習?\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "83bb6165-571d-43d8-8631-f284e19dd9ee",
   "metadata": {},
   "source": [
    "## 初始環境設定"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e234c20e-2f9e-4567-b8bd-7176546fb7b3",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "from pathlib import Path\n",
    "HOME = str(Path.home())\n",
    "Add_Binarry_Path=HOME+'/.local/bin'\n",
    "os.environ['PATH']=os.environ['PATH']+':'+Add_Binarry_Path\n",
    "current_foldr=!pwd\n",
    "current_foldr=current_foldr[0]\n",
    "current_foldr"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e145ec44-0b71-41bd-92c0-fefd9affb806",
   "metadata": {},
   "source": [
    "## 安裝套件"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f26741fe-c66f-407b-ac81-959529bb7b4e",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "## For colab\n",
    "!pip install cohere gdown kaleido langchain openai pyngrok pypdf python-dotenv sentence-transformers tiktoken -q"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "041a9203-2a60-4dad-b22b-895b0a9a1f30",
   "metadata": {},
   "source": [
    "### OPENAI API KEY"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7499fafc-0301-4486-a36d-fb757d2d919b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# OPENAPI KEY method 1\n",
    "\n",
    "!echo \"OPENAI_API_KEY=sk-xxxxxxxxxxx\" > .env\n",
    "from dotenv import load_dotenv\n",
    "load_dotenv() # loads env variables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "45029221-279a-42ec-8e5b-3cbf9cf1a0f2",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "# OPENAPI KEY  method 2\n",
    "\n",
    "import os\n",
    "os.environ[\"OPENAI_API_KEY\"] = \"sk-xxxxxxxxxxx\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c1828b67-676a-4a96-8b15-36ac083647e8",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "# OPENAPI KEY  method 3\n",
    "\n",
    "import os\n",
    "from typing import TextIO\n",
    "from getpass import getpass\n",
    "os.environ[\"OPENAI_API_KEY\"] = getpass()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "131aa253-8cb7-4171-9f32-7923fc3384ea",
   "metadata": {},
   "source": [
    "### LOAD LIBRARY"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4ed4e3bf-77ce-434c-a651-7388d2231277",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "from langchain.llms import OpenAI"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9b644634-295d-4baf-9d66-a1d11fdb3edd",
   "metadata": {},
   "source": [
    "### Load Model\n",
    "temperature 的參數值越小，模型就會回傳越確定的結果。如果調高該參數值，大語言模型可能會返回更隨機的結果，也就是說這可能會帶來更多樣化或更具創造性的產出"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0bb705af-4dca-4fbf-9afc-bc7d0a468606",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "llm = OpenAI(temperature=0.7)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dab5eca6-b96f-4c1d-a7f7-89a99fecbb7e",
   "metadata": {},
   "source": [
    "### QUESTION to Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bc2a2878-965f-4af9-becd-23674f33566d",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "response=llm(\"什麼是聯邦式學習?\")\n",
    "\n",
    "print(response)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.8.16"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}