evaluation/visualization.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import json\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "\n",
    "label1 = 'Fine-tuned'\n",
    "label2 = 'Original'\n",
    "sub_title = 'Mistral 7B_q4_k_m before and after Fine-tuning'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Arithmetics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "dataset = 'artithmetics'\n",
    "\n",
    "# Load data for model 1 from the JSON file\n",
    "with open('finetuned_arithmetic_evalResults/results_2024-09-14T17-38-30.057033.json', 'r') as f:\n",
    "    data_model_1 = json.load(f)\n",
    "with open('original_arithmetic_evalResults/results_2024-09-14T15-11-39.490525.json', 'r') as f:\n",
    "    data_model_2 = json.load(f) \n",
    "\n",
    "# Extract the results\n",
    "results_model_1 = data_model_1['results']\n",
    "results_model_2 = data_model_2['results']\n",
    "\n",
    "# Prepare lists to store the data\n",
    "tasks = []\n",
    "acc_model_1 = []\n",
    "stderr_model_1 = []\n",
    "acc_model_2 = []\n",
    "stderr_model_2 = []\n",
    "\n",
    "# Populate the lists with the relevant data from both models\n",
    "for task, result in results_model_1.items():\n",
    "    tasks.append(task)\n",
    "    acc_model_1.append(result['acc,none'])\n",
    "    stderr_model_1.append(result['acc_stderr,none'])\n",
    "    acc_model_2.append(results_model_2[task]['acc,none'])\n",
    "    stderr_model_2.append(results_model_2[task]['acc_stderr,none'])\n",
    "\n",
    "# Set up the plot\n",
    "fig, ax = plt.subplots(figsize=(10, 6))\n",
    "x = np.arange(len(tasks))\n",
    "\n",
    "# Plot model 1 results\n",
    "ax.bar(x - 0.2, acc_model_1, width=0.4, label=label1, yerr=stderr_model_1, capsize=5)\n",
    "\n",
    "# Plot model 2 results (identical in this case)\n",
    "ax.bar(x + 0.2, acc_model_2, width=0.4, label=label2, yerr=stderr_model_2, capsize=5)\n",
    "\n",
    "# Add labels, title, and grid\n",
    "ax.set_ylabel('Accuracy')\n",
    "ax.set_title(sub_title+f' for {dataset}')\n",
    "ax.set_xticks(x)\n",
    "ax.set_xticklabels(tasks, rotation=45, ha='right')\n",
    "ax.legend()\n",
    "\n",
    "# Show the plot\n",
    "plt.tight_layout()\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### All Accuracies"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import json\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "\n",
    "# Replace these with your actual lists of file paths for both models\n",
    "model_1_files = [\n",
    "    'finetuned_mc_taco_evalResults/results_2024-09-14T18-20-27.796164.json', \n",
    "    'finetuned_arithmetic_evalResults/results_2024-09-14T17-38-30.057033.json',\n",
    "    'finetuned_sciq_evalResults/results_2024-09-14T18-15-30.513794.json',\n",
    "    'finetuned_eval_logiqa_results/results_2024-09-16T17-49-37.154681.json',\n",
    "    'finetuned_eval_codexglue_code2text_results/results_2024-09-16T19-15-42.075638.json'\n",
    "]\n",
    "model_2_files = [\n",
    "    'original_mc_taco_evalResults/unsloth__mistral-7b-bnb-4bit/results_2024-09-14T16-03-38.875262.json', \n",
    "    'original_arithmetic_evalResults/results_2024-09-14T15-11-39.490525.json',\n",
    "    'original_sciq_evalResults/unsloth__mistral-7b-bnb-4bit/results_2024-09-14T15-24-08.081999.json',\n",
    "    'original_logiqa_evalResults/unsloth__mistral-7b-bnb-4bit/results_2024-09-14T15-45-08.855640.json',\n",
    "    'finetuned_eval_codexglue_code2text_results/results_2024-09-16T19-15-42.075638.json'\n",
    "]\n",
    "\n",
    "# Initialize lists to store tasks and accuracy values for both models\n",
    "tasks = []\n",
    "acc_model_1 = []\n",
    "acc_model_2 = []\n",
    "\n",
    "# Function to extract accuracy values from a single JSON file\n",
    "def extract_accuracies(file_path, task_list, acc_list):\n",
    "    with open(file_path, 'r') as f:\n",
    "        data = json.load(f)\n",
    "        results = data.get('results', {})\n",
    "        for task, result in results.items():\n",
    "            if 'acc,none' in result:\n",
    "                # Single accuracy value\n",
    "                task_list.append(task)\n",
    "                acc_list.append(result['acc,none'])\n",
    "            else:\n",
    "                # Handle multiple accuracy types\n",
    "                for key, value in result.items():\n",
    "                    if key.startswith('acc'):\n",
    "                        task_list.append(f\"{task}_{key}\")\n",
    "                        acc_list.append(value)\n",
    "\n",
    "# Extract accuracy data from Model 1 files\n",
    "for file_path in model_1_files:\n",
    "    extract_accuracies(file_path, tasks, acc_model_1)\n",
    "\n",
    "# Extract accuracy data from Model 2 files (ensure the tasks match with Model 1)\n",
    "for file_path in model_2_files:\n",
    "    acc_temp = []\n",
    "    extract_accuracies(file_path, [], acc_temp)\n",
    "    acc_model_2.extend(acc_temp)\n",
    "\n",
    "# Ensure both lists (Model 1 and 2) have the same number of accuracy values\n",
    "assert len(acc_model_1) == len(acc_model_2), \"Mismatch in dataset lengths between Model 1 and Model 2\"\n",
    "\n",
    "# Combine tasks and accuracy values into tuples and sort by Model 1 accuracy (highest to lowest)\n",
    "sorted_data = sorted(zip(tasks, acc_model_1, acc_model_2), key=lambda x: x[1], reverse=True)\n",
    "\n",
    "# Unzip the sorted data\n",
    "tasks_sorted, acc_model_1_sorted, acc_model_2_sorted = zip(*sorted_data)\n",
    "\n",
    "# Set up the plot\n",
    "fig, ax = plt.subplots(figsize=(12, 6))\n",
    "x = np.arange(len(tasks_sorted))\n",
    "\n",
    "# Plot accuracies for Model 1 and Model 2\n",
    "ax.bar(x - 0.2, acc_model_1_sorted, width=0.4, label='Fine-tuned ', capsize=5)\n",
    "ax.bar(x + 0.2, acc_model_2_sorted, width=0.4, label='Original', capsize=5)\n",
    "\n",
    "# Add labels, title, and grid\n",
    "ax.set_ylabel('Accuracy')\n",
    "ax.set_title('Accuracy before and after Fine-Tuning')\n",
    "ax.set_xticks(x)\n",
    "ax.set_xticklabels(tasks_sorted, rotation=45, ha='right')\n",
    "ax.legend()\n",
    "\n",
    "# Show the plot\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Code X Glue"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import json\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "\n",
    "# Load the JSON files for Model 1 and Model 2\n",
    "with open('finetuned_eval_codexglue_code2text_results/results_2024-09-16T19-15-42.075638.json', 'r') as f_model_1, \\\n",
    "     open('original_codexglue_code2text_evalResults/content/original_codexglue_code2text_evalResults/unsloth__mistral-7b-bnb-4bit/results_2024-09-16T23-16-49.567404.json', 'r') as f_model_2:  # Assume model 2 has similar data structure\n",
    "    data_model_1 = json.load(f_model_1)\n",
    "    data_model_2 = json.load(f_model_2)\n",
    "\n",
    "# Extract the results for both models\n",
    "results_model_1 = data_model_1['results']\n",
    "results_model_2 = data_model_2['results']\n",
    "\n",
    "# Prepare lists to store task names and BLEU scores for both models\n",
    "tasks = []\n",
    "bleu_model_1 = []\n",
    "bleu_model_2 = []\n",
    "\n",
    "# Function to extract BLEU scores\n",
    "def extract_bleu_scores(results, task_list, bleu_list):\n",
    "    for task, result in results.items():\n",
    "        if 'smoothed_bleu_4,none' in result:\n",
    "            task_list.append(task)\n",
    "            bleu_list.append(result['smoothed_bleu_4,none'])\n",
    "\n",
    "# Extract BLEU scores for both models\n",
    "extract_bleu_scores(results_model_1, tasks, bleu_model_1)\n",
    "extract_bleu_scores(results_model_2, [], bleu_model_2)  # Assuming tasks are the same\n",
    "\n",
    "# Combine tasks and BLEU scores into tuples and sort by Model 1 BLEU scores (from highest to lowest)\n",
    "sorted_data = sorted(zip(tasks, bleu_model_1, bleu_model_2), key=lambda x: x[1], reverse=True)\n",
    "\n",
    "# Unzip the sorted data\n",
    "tasks_sorted, bleu_model_1_sorted, bleu_model_2_sorted = zip(*sorted_data)\n",
    "\n",
    "# Set up the plot\n",
    "fig, ax = plt.subplots(figsize=(12, 6))\n",
    "x = np.arange(len(tasks_sorted))\n",
    "\n",
    "# Plot BLEU scores for Model 1 and Model 2\n",
    "ax.bar(x - 0.2, bleu_model_1_sorted, width=0.4, label='Fine-tuned', capsize=5)\n",
    "ax.bar(x + 0.2, bleu_model_2_sorted, width=0.4, label='Original', capsize=5)\n",
    "\n",
    "# Add labels, title, and grid\n",
    "ax.set_ylabel('BLEU Score')\n",
    "ax.set_title('Performance on CodeXGlue before and after Fine-Tuning')\n",
    "ax.set_xticks(x)\n",
    "ax.set_xticklabels(tasks_sorted, rotation=45, ha='right')\n",
    "ax.legend()\n",
    "\n",
    "# Show the plot\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "robotbek",
   "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
}