README.md

metadata

title: Mistral Fine-tuned Model
emoji: 🤖
colorFrom: blue
colorTo: purple
sdk: docker
app_port: 7860

🤖 Mistral Fine-tuned Model

Flask API with separate HTML/CSS/JS frontend for KASHH-4/mistral_fine-tuned model.

🚀 What This Is

A Flask API server with separate frontend files:

• Backend: Python Flask with CORS
• Frontend: HTML + CSS + JavaScript
• Clean separation of concerns
• API-first design

📁 Project Structure

e:\EDI\hf-node-app\
├── app.py              # Main Gradio application
├── requirements.txt    # Python dependencies
├── README.md          # This file
└── .gitignore         # Git ignore rules

🔧 Deploy to Hugging Face Spaces

Step 1: Create a Space

1. Go to https://huggingface.co/spaces
2. Click "Create new Space"
3. Configure:
   • Owner: KASHH-4 (or your account)
   • Space name: mistral-api (or any name)
   • SDK: Gradio
   • Hardware: CPU basic (Free)
   • Visibility: Public
4. Click "Create Space"

Step 2: Upload Files

Upload these 3 files to your Space:

• app.py
• requirements.txt
• README.md (optional)

Via Web UI:

1. Click "Files" tab
2. Click "Add file" → "Upload files"
3. Drag and drop the files
4. Commit changes

Via Git:

git init
git remote add origin https://huggingface.co/spaces/KASHH-4/mistral-api
git add app.py requirements.txt README.md .gitignore
git commit -m "Initial deployment"
git push origin main

Step 3: Wait for Deployment

• First build takes 5-10 minutes
• Watch the logs for "Running on..."
• Your Space will be live at: https://kashh-4-mistral-api.hf.space

🧪 Test Your Space

Web Interface

Visit: https://huggingface.co/spaces/KASHH-4/mistral-api

API Endpoint

curl -X POST "https://kashh-4-mistral-api.hf.space/api/predict" \
  -H "Content-Type: application/json" \
  -d '{"data":["Hello, how are you?"]}'

From JavaScript/Node.js

const response = await fetch('https://kashh-4-mistral-api.hf.space/api/predict', {
  method: 'POST',
  headers: { 'Content-Type': 'application/json' },
  body: JSON.stringify({ data: ["Your prompt here"] })
});

const result = await response.json();
console.log(result.data[0]); // Generated text

From Python

import requests

response = requests.post(
    'https://kashh-4-mistral-api.hf.space/api/predict',
    json={'data': ['Your prompt here']}
)

print(response.json()['data'][0])

💰 Cost

100% FREE on HF Spaces:

• Free CPU tier (slower, ~10-30 sec per request)
• Sleeps after 48h inactivity (30 sec wake-up)
• Perfect for demos, personal projects, testing

Optional Upgrades:

• GPU T4 Small: $0.60/hour (much faster, 2-5 sec)
• GPU A10G: $3.15/hour (very fast, 1-2 sec)

Upgrade in: Space Settings → Hardware

🔧 Local Testing (Optional)

If you have Python installed and want to test locally before deploying:

# Install dependencies
pip install -r requirements.txt

# Run locally
python app.py

# Visit: http://localhost:7860

Requirements:

• Python 3.9+
• 16GB+ RAM (for model loading)
• GPU recommended but not required

📋 Model Configuration

The app is configured for KASHH-4/mistral_fine-tuned. To use a different model, edit app.py:

MODEL_NAME = "your-org/your-model"

🆘 Troubleshooting

Space stuck on "Building":

• Check logs for errors
• Model might be too large for free CPU
• Try: Restart Space in Settings

Space shows "Runtime Error":

• Check if model exists and is public
• Verify model format is compatible with transformers
• Try smaller model first to test

Slow responses:

• Normal on free CPU tier
• Upgrade to GPU for faster inference
• Or use smaller model

📞 Support

Issues? Check the deployment guide in huggingface-space/DEPLOYMENT-GUIDE.md

---

🗑️ Cleanup Old Files

If you followed earlier Node.js instructions, delete unnecessary files:

See CLEANUP.md for full list of files to remove.

License

MIT