evaluate_model.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Evaluating models based on medQA dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Using the https://huggingface.co/datasets/medmcqa/viewer/default/validation?row=0 dataset we asses the models performance on the medQA dataset."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Libraries & Setup"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%pip install datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%pip install transformers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%pip install pretty-errors"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/homebrew/anaconda3/envs/adl/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"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "import seaborn as sns\n",
    "\n",
    "from datasets import load_dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14/12/2023 13:42:10] - INFO : Found credentials in shared credentials file: ~/.aws/credentials\n",
      "sagemaker.config INFO - Not applying SDK defaults from location: /opt/homebrew/share/sagemaker/config.yaml\n",
      "sagemaker.config INFO - Not applying SDK defaults from location: /Users/tilmankerl/Library/Application Support/sagemaker/config.yaml\n"
     ]
    }
   ],
   "source": [
    "from chat_doc.inference.deploy_endpoint import SageMakerDeployment\n",
    "from chat_doc.inference.chat import Chat\n",
    "from chat_doc.config import SEED"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model loading"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "s3_model_uri = \"s3://sagemaker-eu-central-1-228610994900/huggingface-qlora-2023-12-08-15-01-13-2023-12-08-15-01-14-300/output/model.tar.gz\"\n",
    "deployment = SageMakerDeployment(s3_model_uri)\n",
    "llm_endpoint = deployment.deploy_model()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Setup"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "chat = Chat(llm_endpoint)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data Loading"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset = load_dataset(\"medmcqa\")\n",
    "validation = dataset[\"validation\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Evaluation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Filter: 100%|██████████| 4183/4183 [00:00<00:00, 106575.47 examples/s]\n"
     ]
    }
   ],
   "source": [
    "# filter huggingface dataset to only include single choice questions\n",
    "validation = validation.filter(lambda example: example[\"choice_type\"] == \"single\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "validation_samples = validation.to_pandas().sample(100, random_state=SEED)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def build_qa_query(row):\n",
    "    return f\"\"\"Please answer the {row['choice_type']}-choice question to the best of your knowledge by just returning the correct option. The subject is {row['subject_name']}.\n",
    "\n",
    "    Question: {row['question']}\n",
    "    Options: \n",
    "    A) {row['opa']}\n",
    "    B) {row['opb']}\n",
    "    C) {row['opc']}\n",
    "    D) {row['opd']}\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "validation_samples = [\n",
    "    chat.predict(\n",
    "        build_qa_query(row)\n",
    "    ) for _, row in validation_samples.iterrows()\n",
    "\n",
    "]"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "adl",
   "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": "3.11.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}