---
language: en
tags:
  - e-commerce
  - intent-classification
  - chatbot
  - pytorch
  - gradio
license: mit
---

# 🛍️ EasySell AI — E-Commerce Chatbot SLM

A **100M parameter** GPT-style Small Language Model trained for e-commerce 
intent classification and tool-calling agent for online shopping.

## Model Details
- **Architecture:** Custom GPT (14 layers, 12 heads, 768 dim)
- **Parameters:** ~100M
- **Trained on:** 30,000 synthetic e-commerce conversations
- **Intents:** GET_LINK, SEARCH_PRODUCT, ADD_TO_CART, CHECKOUT, FAQ, CLARIFY, REFUSE

## Features
- 🛒 Add to cart with quantity control
- 📦 Order tracking, cancel & refund workflows  
- 🧾 Invoice generation with GST
- 💳 Razorpay payment integration
- 🎟️ Coupon system
- 📊 Product comparison
- 🤖 50+ FAQ auto-answers
- 👨‍💼 Human agent routing

## Usage
```python
import torch, json
from huggingface_hub import hf_hub_download

# Download model
model_path = hf_hub_download(repo_id="YOUR_USERNAME/easysell-ai", filename="model.pt")
config_path = hf_hub_download(repo_id="YOUR_USERNAME/easysell-ai", filename="config.json")

# Load config
with open(config_path) as f:
    cfg = json.load(f)

# Load model (after defining your GPT class)
model.load_state_dict(torch.load(model_path, map_location="cpu"))
model.eval()
```

## Demo
Run the Gradio UI using `GRADIO_FINAL.py` after loading the model.

## Files
| File | Description |
|------|-------------|
| `model.pt` | Model weights |
| `config.json` | Model architecture config |
| `tokenizer.json` | BPE tokenizer |
| `products_description.json` | Product catalog |
| `products_images.json` | Product image URLs |
| `CELL_AGENT_FINAL.py` | Tool-calling agent |
| `GRADIO_FINAL.py` | Gradio chat UI |