461 lines
61 KiB
Plaintext
461 lines
61 KiB
Plaintext
{
|
||
"cells": [
|
||
{
|
||
"cell_type": "markdown",
|
||
"id": "db8736a7-ed94-441c-9556-831fa57b5a10",
|
||
"metadata": {},
|
||
"source": [
|
||
"# The Product Pricer Fine Tuning\n",
|
||
"\n",
|
||
"Submitted By: Bharat Puri\n",
|
||
"\n",
|
||
"A model that can estimate how much something costs, from its description.\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 5,
|
||
"id": "681c717b-4c24-4ac3-a5f3-3c5881d6e70a",
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": [
|
||
"# imports\n",
|
||
"\n",
|
||
"import os\n",
|
||
"import re\n",
|
||
"import math\n",
|
||
"import json\n",
|
||
"import random\n",
|
||
"from dotenv import load_dotenv\n",
|
||
"from huggingface_hub import login\n",
|
||
"import matplotlib.pyplot as plt\n",
|
||
"import pandas as pd\n",
|
||
"import numpy as np\n",
|
||
"import pickle\n",
|
||
"from collections import Counter\n",
|
||
"import sys\n",
|
||
"sys.path.append(os.path.abspath(os.path.join(\"..\", \"..\"))) \n",
|
||
"from openai import OpenAI\n",
|
||
"from anthropic import Anthropic\n",
|
||
"from sklearn.model_selection import train_test_split\n",
|
||
"from sklearn.metrics import mean_absolute_error\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 2,
|
||
"id": "36d05bdc-0155-4c72-a7ee-aa4e614ffd3c",
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": [
|
||
"# environment\n",
|
||
"\n",
|
||
"load_dotenv(override=True)\n",
|
||
"os.environ['OPENAI_API_KEY'] = os.getenv('OPENAI_API_KEY', 'your-key-if-not-using-env')\n",
|
||
"os.environ['ANTHROPIC_API_KEY'] = os.getenv('ANTHROPIC_API_KEY', 'your-key-if-not-using-env')\n",
|
||
"os.environ['HF_TOKEN'] = os.getenv('HF_TOKEN', 'your-key-if-not-using-env')"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 3,
|
||
"id": "4dd3aad2-6f99-433c-8792-e461d2f06622",
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"name": "stderr",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured.\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"# Log in to HuggingFace\n",
|
||
"\n",
|
||
"hf_token = os.environ['HF_TOKEN']\n",
|
||
"login(hf_token, add_to_git_credential=True)"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 7,
|
||
"id": "884a50bd-8cae-425e-8e56-f079fc3e65ce",
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"Input columns: [\"How much does this cost to the nearest dollar?\\n\\nOEM AC Compressor w/A/C Repair Kit For Ford F150 F-150 V8 & Lincoln Mark LT 2007 2008 - BuyAutoParts NEW\\nAs one of the world's largest automotive parts suppliers, our parts are trusted every day by mechanics and vehicle owners worldwide. This A/C Compressor and Components Kit is manufactured and tested to the strictest OE standards for unparalleled performance. Built for trouble-free ownership and 100% visually inspected and quality tested, this A/C Compressor and Components Kit is backed by our 100% satisfaction guarantee. Guaranteed Exact Fit for easy installation 100% BRAND NEW, premium ISO/TS 16949 quality - tested to meet or exceed OEM specifications Engineered for superior durability, backed by industry-leading unlimited-mileage warranty Included in this K\\n\\nPrice is $\", '0']\n",
|
||
"Output columns: [\"How much does this cost to the nearest dollar?\\n\\nOEM AC Compressor w/A/C Repair Kit For Ford F150 F-150 V8 & Lincoln Mark LT 2007 2008 - BuyAutoParts NEW\\nAs one of the world's largest automotive parts suppliers, our parts are trusted every day by mechanics and vehicle owners worldwide. This A/C Compressor and Components Kit is manufactured and tested to the strictest OE standards for unparalleled performance. Built for trouble-free ownership and 100% visually inspected and quality tested, this A/C Compressor and Components Kit is backed by our 100% satisfaction guarantee. Guaranteed Exact Fit for easy installation 100% BRAND NEW, premium ISO/TS 16949 quality - tested to meet or exceed OEM specifications Engineered for superior durability, backed by industry-leading unlimited-mileage warranty Included in this K\\n\\nPrice is $\", '120']\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"# =============================================\n",
|
||
"# Step 1 – Load and Inspect Dataset (CSV files)\n",
|
||
"# =============================================\n",
|
||
"\n",
|
||
"df_input = pd.read_csv(\"../../human_input.csv\")\n",
|
||
"df_output = pd.read_csv(\"../../human_output.csv\")\n",
|
||
"\n",
|
||
"print(\"Input columns:\", df_input.columns.tolist())\n",
|
||
"print(\"Output columns:\", df_output.columns.tolist())\n",
|
||
"\n",
|
||
"# Detect correct column names automatically\n",
|
||
"input_col = df_input.columns[0] # first column name\n",
|
||
"output_col = df_output.columns[0] # first column name\n",
|
||
"\n",
|
||
"data = pd.DataFrame({\n",
|
||
" \"prompt\": df_input[input_col].astype(str),\n",
|
||
" \"completion\": df_output[output_col].astype(str)\n",
|
||
"})"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 8,
|
||
"id": "b0a6fb86-74a4-403c-ab25-6db2d74e9d2b",
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"Training samples: 199 | Validation samples: 50\n",
|
||
"✅ Train and validation data prepared successfully.\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"# =============================================\n",
|
||
"# Step 2 – Split into Train and Validation Sets\n",
|
||
"# =============================================\n",
|
||
"\n",
|
||
"from sklearn.model_selection import train_test_split\n",
|
||
"\n",
|
||
"# Keep this small to minimize cost\n",
|
||
"train_df, val_df = train_test_split(data, test_size=0.2, random_state=42)\n",
|
||
"\n",
|
||
"print(f\"Training samples: {len(train_df)} | Validation samples: {len(val_df)}\")\n",
|
||
"\n",
|
||
"# Save to JSONL format (required by OpenAI fine-tuning API)\n",
|
||
"train_df.to_json(\"train.jsonl\", orient=\"records\", lines=True)\n",
|
||
"val_df.to_json(\"val.jsonl\", orient=\"records\", lines=True)\n",
|
||
"\n",
|
||
"print(\"✅ Train and validation data prepared successfully.\")"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 9,
|
||
"id": "c830ed3e-24ee-4af6-a07b-a1bfdcd39278",
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"data": {
|
||
"text/html": [
|
||
"<div>\n",
|
||
"<style scoped>\n",
|
||
" .dataframe tbody tr th:only-of-type {\n",
|
||
" vertical-align: middle;\n",
|
||
" }\n",
|
||
"\n",
|
||
" .dataframe tbody tr th {\n",
|
||
" vertical-align: top;\n",
|
||
" }\n",
|
||
"\n",
|
||
" .dataframe thead th {\n",
|
||
" text-align: right;\n",
|
||
" }\n",
|
||
"</style>\n",
|
||
"<table border=\"1\" class=\"dataframe\">\n",
|
||
" <thead>\n",
|
||
" <tr style=\"text-align: right;\">\n",
|
||
" <th></th>\n",
|
||
" <th>prompt</th>\n",
|
||
" <th>completion</th>\n",
|
||
" </tr>\n",
|
||
" </thead>\n",
|
||
" <tbody>\n",
|
||
" <tr>\n",
|
||
" <th>137</th>\n",
|
||
" <td>How much does this cost to the nearest dollar?...</td>\n",
|
||
" <td>How much does this cost to the nearest dollar?...</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>6</th>\n",
|
||
" <td>How much does this cost to the nearest dollar?...</td>\n",
|
||
" <td>How much does this cost to the nearest dollar?...</td>\n",
|
||
" </tr>\n",
|
||
" <tr>\n",
|
||
" <th>97</th>\n",
|
||
" <td>How much does this cost to the nearest dollar?...</td>\n",
|
||
" <td>How much does this cost to the nearest dollar?...</td>\n",
|
||
" </tr>\n",
|
||
" </tbody>\n",
|
||
"</table>\n",
|
||
"</div>"
|
||
],
|
||
"text/plain": [
|
||
" prompt \\\n",
|
||
"137 How much does this cost to the nearest dollar?... \n",
|
||
"6 How much does this cost to the nearest dollar?... \n",
|
||
"97 How much does this cost to the nearest dollar?... \n",
|
||
"\n",
|
||
" completion \n",
|
||
"137 How much does this cost to the nearest dollar?... \n",
|
||
"6 How much does this cost to the nearest dollar?... \n",
|
||
"97 How much does this cost to the nearest dollar?... "
|
||
]
|
||
},
|
||
"execution_count": 9,
|
||
"metadata": {},
|
||
"output_type": "execute_result"
|
||
}
|
||
],
|
||
"source": [
|
||
"train_df.head(3)\n",
|
||
"val_df.head(3)\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": null,
|
||
"id": "5c9b05f4-c9eb-462c-8d86-de9140a2d985",
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": [
|
||
"# =============================================\n",
|
||
"# Step 3 – Define Fine-Tuning Configuration\n",
|
||
"# =============================================\n",
|
||
"\n",
|
||
"hyperparams = {\n",
|
||
" \"model\": \"gpt-4o-mini\", \n",
|
||
" \"n_epochs\": 1, \n",
|
||
" \"batch_size\": 4, # Small batch = less token use\n",
|
||
" \"learning_rate_multiplier\": 0.5, # Gentle learning rate\n",
|
||
" \"suffix\": \"week6_lowcost_bharat\" # Custom suffix for tracking\n",
|
||
"}\n",
|
||
"\n",
|
||
"print(\"✅ Fine-tuning configuration defined:\")\n",
|
||
"for k, v in hyperparams.items():\n",
|
||
" print(f\"{k:25}: {v}\")\n"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": null,
|
||
"id": "e8367135-f40e-43e1-8f3c-09e990ab1194",
|
||
"metadata": {},
|
||
"outputs": [],
|
||
"source": [
|
||
"# OpenAI recommends fine-tuning with populations of 50-100 examples\n",
|
||
"# But as our examples are very small, I'm suggesting we go with 200 examples (and 1 epoch)\n",
|
||
"\n",
|
||
"fine_tune_train = train[:200]\n",
|
||
"fine_tune_validation = train[200:250]"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 11,
|
||
"id": "8ae2fb3c-1cff-4ce3-911e-627c970edd7b",
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"\n",
|
||
"⚙️ Simulating fine-tuning process (no API cost)...\n",
|
||
"Epoch 1/1 training...\n",
|
||
"Fine-tuning complete ✅ (simulated)\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"# =============================================\n",
|
||
"# Step 4 – Launch Fine-Tuning Job or Simulate\n",
|
||
"# =============================================\n",
|
||
"\n",
|
||
"import time\n",
|
||
"from openai import OpenAI\n",
|
||
"\n",
|
||
"# Initialize the OpenAI client\n",
|
||
"client = OpenAI(api_key=os.getenv(\"OPENAI_API_KEY\"))\n",
|
||
"\n",
|
||
"# Toggle this flag to switch between simulation and real fine-tuning\n",
|
||
"simulate = True # ✅ Default: Free simulation mode\n",
|
||
"\n",
|
||
"if simulate:\n",
|
||
" print(\"\\n⚙️ Simulating fine-tuning process (no API cost)...\")\n",
|
||
" for i in range(hyperparams['n_epochs']):\n",
|
||
" print(f\"Epoch {i+1}/{hyperparams['n_epochs']} training...\")\n",
|
||
" time.sleep(1)\n",
|
||
" print(\"Fine-tuning complete ✅ (simulated)\")\n",
|
||
"else:\n",
|
||
" print(\"\\n🚀 Launching real fine-tuning job...\")\n",
|
||
"\n",
|
||
" # Upload train and validation files\n",
|
||
" train_file = client.files.create(file=open(\"train.jsonl\", \"rb\"), purpose=\"fine-tune\")\n",
|
||
" val_file = client.files.create(file=open(\"val.jsonl\", \"rb\"), purpose=\"fine-tune\")\n",
|
||
"\n",
|
||
" # Create fine-tuning job\n",
|
||
" job = client.fine_tuning.jobs.create(\n",
|
||
" training_file=train_file.id,\n",
|
||
" validation_file=val_file.id,\n",
|
||
" **hyperparams\n",
|
||
" )\n",
|
||
"\n",
|
||
" print(\"✅ Fine-tuning job created successfully!\")\n",
|
||
" print(\"Job ID:\", job.id)"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 12,
|
||
"id": "1aa280f6-1227-426a-a2e2-1ce985feba1e",
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"\n",
|
||
"🔍 Evaluating model performance...\n",
|
||
"✅ Simulation mode: generated random prediction values for evaluation.\n",
|
||
"\n",
|
||
"📊 Validation Mean Absolute Error (MAE): 3.30\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"# =============================================\n",
|
||
"# Step 5 – Evaluate Fine-Tuned (or Simulated) Model\n",
|
||
"# =============================================\n",
|
||
"\n",
|
||
"from sklearn.metrics import mean_absolute_error\n",
|
||
"import numpy as np\n",
|
||
"\n",
|
||
"print(\"\\n🔍 Evaluating model performance...\")\n",
|
||
"\n",
|
||
"# Keep evaluation small to minimize cost\n",
|
||
"val_df = val_df.head(5)\n",
|
||
"\n",
|
||
"predictions = []\n",
|
||
"actuals = []\n",
|
||
"\n",
|
||
"if simulate:\n",
|
||
" # Simulated predictions for free mode\n",
|
||
" predictions = np.random.uniform(70, 90, len(val_df))\n",
|
||
" actuals = np.random.uniform(70, 90, len(val_df))\n",
|
||
" print(\"✅ Simulation mode: generated random prediction values for evaluation.\")\n",
|
||
"else:\n",
|
||
" # Real evaluation using fine-tuned model\n",
|
||
" print(\"🧠 Generating predictions using fine-tuned model...\")\n",
|
||
" for _, row in val_df.iterrows():\n",
|
||
" response = client.chat.completions.create(\n",
|
||
" model=f\"ft:{hyperparams['model']}:{hyperparams['suffix']}\",\n",
|
||
" messages=[{\"role\": \"user\", \"content\": row['prompt']}],\n",
|
||
" )\n",
|
||
" pred = response.choices[0].message.content.strip()\n",
|
||
" predictions.append(pred)\n",
|
||
" actuals.append(row['completion'])\n",
|
||
"\n",
|
||
"# Try calculating MAE if numeric outputs\n",
|
||
"try:\n",
|
||
" preds_float = [float(p) for p in predictions]\n",
|
||
" acts_float = [float(a) for a in actuals]\n",
|
||
" mae = mean_absolute_error(acts_float, preds_float)\n",
|
||
" print(f\"\\n📊 Validation Mean Absolute Error (MAE): {mae:.2f}\")\n",
|
||
"except:\n",
|
||
" print(\"\\n⚠️ Non-numeric outputs detected — qualitative comparison recommended.\")\n",
|
||
" for i in range(len(val_df)):\n",
|
||
" print(f\"\\nPrompt: {val_df.iloc[i]['prompt']}\")\n",
|
||
" print(f\"→ Prediction: {predictions[i]}\")\n",
|
||
" print(f\"→ Actual: {actuals[i]}\")"
|
||
]
|
||
},
|
||
{
|
||
"cell_type": "code",
|
||
"execution_count": 16,
|
||
"id": "c0e5b56c-8a0b-4d8e-a112-ce87efb4e152",
|
||
"metadata": {},
|
||
"outputs": [
|
||
{
|
||
"data": {
|
||
"image/png": "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",
|
||
"text/plain": [
|
||
"<Figure size 600x400 with 1 Axes>"
|
||
]
|
||
},
|
||
"metadata": {},
|
||
"output_type": "display_data"
|
||
},
|
||
{
|
||
"name": "stdout",
|
||
"output_type": "stream",
|
||
"text": [
|
||
"\n",
|
||
"===== WEEK 6 REFLECTION =====\n",
|
||
"✅ Completed the full fine-tuning workflow successfully.\n",
|
||
"🧠 Simulation mode enabled full understanding without any API cost.\n",
|
||
"📊 Validation MAE: 3.30 (simulated)\n",
|
||
"🔍 Learned how to prepare data, configure fine-tuning, and evaluate models safely.\n",
|
||
"💡 Next step: Try real fine-tuning (simulate=False) on small data if free credits are available.\n"
|
||
]
|
||
}
|
||
],
|
||
"source": [
|
||
"# =============================================\n",
|
||
"# Step 6 – Visualize and Reflect (Fixed)\n",
|
||
"# =============================================\n",
|
||
"\n",
|
||
"import matplotlib.pyplot as plt\n",
|
||
"\n",
|
||
"# Plot simulated predictions vs actuals\n",
|
||
"plt.figure(figsize=(6, 4))\n",
|
||
"plt.plot(preds_float, label=\"Predicted\", marker='o')\n",
|
||
"plt.plot(acts_float, label=\"Actual\", marker='x')\n",
|
||
"plt.title(\"Validation Predictions vs Actuals (Simulated)\")\n",
|
||
"plt.xlabel(\"Sample Index\")\n",
|
||
"plt.ylabel(\"Value\")\n",
|
||
"plt.legend()\n",
|
||
"plt.grid(True)\n",
|
||
"plt.show()\n",
|
||
"\n",
|
||
"# Summary Reflection\n",
|
||
"print(\"\\n===== WEEK 6 REFLECTION =====\")\n",
|
||
"print(\"✅ Completed the full fine-tuning workflow successfully.\")\n",
|
||
"print(\"🧠 Simulation mode enabled full understanding without any API cost.\")\n",
|
||
"print(\"📊 Validation MAE: 3.30 (simulated)\")\n",
|
||
"print(\"🔍 Learned how to prepare data, configure fine-tuning, and evaluate models safely.\")\n",
|
||
"print(\"💡 Next step: Try real fine-tuning (simulate=False) on small data if free credits are available.\")\n"
|
||
]
|
||
}
|
||
],
|
||
"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": "3.11.14"
|
||
}
|
||
},
|
||
"nbformat": 4,
|
||
"nbformat_minor": 5
|
||
}
|