langchain_local_using.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "4e7a48be-d8f7-4443-b9c6-c1d728fbc8e1",
   "metadata": {},
   "outputs": [],
   "source": [
    "from langchain import HuggingFacePipeline, PromptTemplate, LLMChain\n",
    "from langchain.prompts import (\n",
    "    ChatPromptTemplate,\n",
    "    SystemMessagePromptTemplate,\n",
    "    HumanMessagePromptTemplate,\n",
    ")\n",
    "from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline\n",
    "import torch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "50bcc067-a43a-4a6e-8a41-158c8e7efc10",
   "metadata": {},
   "outputs": [],
   "source": [
    "# 基本パラメータ\n",
    "model_name = \"elyza/ELYZA-japanese-Llama-2-7b-instruct\"\n",
    "task = \"text-generation\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "7c0b2818-3f10-40ac-99d1-40b4e0a7fe9d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "device(type='cuda')"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# GPUの確認\n",
    "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
    "device"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "568f42e6-485c-4d16-8946-b5eeed8f2d57",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "7420d1901cd6496db92701d74f7dac62",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Loading checkpoint shards:   0%|          | 0/2 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# モデルのダウンロード\n",
    "tokenizer = AutoTokenizer.from_pretrained(model_name)\n",
    "model = AutoModelForCausalLM.from_pretrained(model_name).to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "672151dc-5780-4175-8635-b190fa8e9fb3",
   "metadata": {},
   "outputs": [],
   "source": [
    "# LLMs: langchainで上記モデルを利用する\n",
    "pipe = pipeline(\n",
    "    task, \n",
    "    model=model,\n",
    "    tokenizer=tokenizer,\n",
    "    device=0,            # GPUを使うことを指定 (cuda:0と同義)\n",
    "    framework='pt',      # モデルをPyTorchで読み込むことを指定\n",
    "    max_new_tokens=1024,\n",
    "    #temperature=0.1,\n",
    ")\n",
    "\n",
    "llm = HuggingFacePipeline(pipeline=pipe)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "4a638e89-5526-47e4-b5eb-c2c5c6a4dad6",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Prompts: プロンプトを作成\n",
    "template = \"<s>[INST] <<SYS>>あなたはユーザの質問に回答する優秀なアシスタントです。以下の質問に可能な限り丁寧に回答してください。 <</SYS>>\\n\\n{question}[/INST]\"\n",
    "prompt = PromptTemplate.from_template(template)\n",
    "#sys_prompt = SystemMessagePromptTemplate.from_template(\"あなたはユーザの質問に回答する優秀なアシスタントです。以下の質問に可能な限り丁寧に回答してください。\")\n",
    "#hum_prompt =  HumanMessagePromptTemplate.from_template(\"{question}\")\n",
    "#prompt = ChatPromptTemplate.from_messages([sys_prompt, hum_prompt])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "65b6708e-4633-413f-96b2-fea3a7579937",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Chains: llmを利用可能な状態にする\n",
    "llm_chain = LLMChain(prompt=prompt, llm=llm, verbose=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "b2af34d3-768a-46aa-87ba-1dd33dba1bb2",
   "metadata": {},
   "outputs": [],
   "source": [
    "question = \"カレーライスとは何ですか？\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "58db6470-64fc-4ad2-b1f3-e2c4a91fc017",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "\n",
      "\u001b[1m> Entering new LLMChain chain...\u001b[0m\n",
      "Prompt after formatting:\n",
      "\u001b[32;1m\u001b[1;3m<s>[INST] <<SYS>>あなたはユーザの質問に回答する優秀なアシスタントです。以下の質問に可能な限り丁寧に回答してください。 <</SYS>>\n",
      "\n",
      "カレーライスとは何ですか？[/INST]\u001b[0m\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/user/.pyenv/versions/3.10.12/lib/python3.10/site-packages/transformers/generation/utils.py:1411: UserWarning: You have modified the pretrained model configuration to control generation. This is a deprecated strategy to control generation and will be removed soon, in a future version. Please use a generation configuration file (see https://huggingface.co/docs/transformers/main_classes/text_generation )\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "\u001b[1m> Finished chain.\u001b[0m\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'  カレーライスとは、ご飯の上にカレーが乗っている料理のことです。一般的には、米を炊いた後、その米をカレーのルーで炒めたり、浸けたりして作ります。また、具材としてチキンやエビ、肉などが入ったり、サラダやヨーグルトが添えられることもあります。'"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "llm_chain.run({\"question\":question})"
   ]
  }
 ],
 "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": "3.10.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}