Update README.md

README.md

@@ -1,252 +1,37 @@
-# 🎭 Emotion Recognition API
-
-A FastAPI-based emotion recognition system using HuBERT (Hidden-Unit BERT) for audio emotion classification.
-
-## 📋 Features
-
-- **Real-time Emotion Detection**: Analyze audio files and detect emotions
-- **Multiple Format Support**: WAV, MP3, FLAC, OGG, M4A
-- **Batch Processing**: Process multiple audio files at once
-- **RESTful API**: Easy integration with any application
-- **High Accuracy**: Fine-tuned HuBERT model for emotion classification
-
-## 🎯 Supported Emotions
-
-- Angry/Disgust
-- Happy/Surprised
-- Neutral/Calm
-- Sad/Fearful
-
-## 🚀 Quick Start
-
-### Using the API
-
-1. **Single Prediction**
-```bash
-curl -X POST "http://your-space-url/predict" \
-  -F "file=@your_audio.wav"
-```
-
-2. **Batch Prediction**
-```bash
-curl -X POST "http://your-space-url/predict_batch" \
-  -F "files=@audio1.wav" \
-  -F "files=@audio2.wav"
-```
-
-3. **Get Available Labels**
-```bash
-curl "http://your-space-url/labels"
-```
+---
+title: HuBERT Emotion Recognition
+emoji: 🎧
+colorFrom: blue
+colorTo: purple
+sdk: docker
+app_port: 7860
+---
 
-4. **Health Check**
-```bash
-curl "http://your-space-url/health"
-```
+## 🎧 HuBERT Emotion Recognition API
 
-## 📖 API Documentation
+This Space provides an emotion recognition API for speech audio using **HuBERT**.
 
-Once deployed, visit `/docs` for interactive API documentation (Swagger UI).
+### 🎯 Supported emotions
+- Neutral / Calm
+- Happy / Surprised
+- Angry / Disgust
+- Sad / Fearful
 
-### Endpoints
+### 🚀 API Endpoint
 
-#### `POST /predict`
-Upload a single audio file for emotion prediction.
+**POST** `/predict`
 
-**Request:**
-- Form data with `file` parameter (audio file)
+Upload a `.wav` file.
 
-**Response:**
+### 📦 Response
 ```json
 {
-  "success": true,
-  "predicted_emotion": "Happy/Surprised",
-  "confidence": 0.8542,
-  "all_probabilities": {
-    "Angry/Disgust": 0.0234,
-    "Happy/Surprised": 0.8542,
-    "Neutral/Calm": 0.0891,
-    "Sad/Fearful": 0.0333
-  },
-  "filename": "sample.wav"
+  "emotion": "Happy/Surprised",
+  "confidence": 0.87,
+  "probabilities": {
+    "Happy/Surprised": 0.87,
+    "Neutral/Calm": 0.05,
+    "Angry/Disgust": 0.04,
+    "Sad/Fearful": 0.04
+  }
 }
-```
-
-#### `POST /predict_batch`
-Upload multiple audio files (max 10) for batch prediction.
-
-**Request:**
-- Form data with multiple `files` parameters
-
-**Response:**
-```json
-{
-  "success": true,
-  "results": [
-    {
-      "filename": "audio1.wav",
-      "predicted_emotion": "Happy/Surprised",
-      "confidence": 0.8542
-    },
-    {
-      "filename": "audio2.wav",
-      "predicted_emotion": "Sad/Fearful",
-      "confidence": 0.7231
-    }
-  ],
-  "total_files": 2
-}
-```
-
-#### `GET /labels`
-Get all available emotion labels.
-
-#### `GET /health`
-Check API health status.
-
-## 🔧 Setup Instructions
-
-### Prerequisites
-- Python 3.10+
-- Your trained HuBERT model files
-
-### Local Development
-
-1. **Clone the repository**
-```bash
-git clone <your-repo>
-cd <repo-name>
-```
-
-2. **Install dependencies**
-```bash
-pip install -r requirements.txt
-```
-
-3. **Add your model**
-Place your trained model files in the `model/` directory:
-```
-model/
-├── config.json
-├── preprocessor_config.json
-├── pytorch_model.bin
-└── (other model files)
-```
-
-4. **Run the server**
-```bash
-uvicorn app:app --host 0.0.0.0 --port 7860
-```
-
-5. **Test the API**
-Visit `http://localhost:7860/docs` for interactive documentation.
-
-### Deploying to Hugging Face Spaces
-
-1. **Create a new Space**
-   - Go to [Hugging Face Spaces](https://huggingface.co/spaces)
-   - Click "Create new Space"
-   - Choose "Docker" as the SDK
-   - Name your Space
-
-2. **Upload files**
-   Upload the following files to your Space:
-   - `app.py`
-   - `requirements.txt`
-   - `Dockerfile`
-   - `README.md`
-   - Your `model/` directory with all model files
-
-3. **Configure Space**
-   - The Space will automatically build using the Dockerfile
-   - Once built, your API will be available at `https://your-username-space-name.hf.space`
-
-## 📦 Model Files Required
-
-Make sure your `model/` directory contains:
-- `config.json` - Model configuration
-- `preprocessor_config.json` - Feature extractor configuration
-- `pytorch_model.bin` - Model weights
-- Any other files saved by `save_pretrained()`
-
-## 🐍 Python Client Example
-
-```python
-import requests
-
-# Predict emotion from audio file
-url = "http://your-space-url/predict"
-files = {"file": open("audio.wav", "rb")}
-response = requests.post(url, files=files)
-result = response.json()
-
-print(f"Emotion: {result['predicted_emotion']}")
-print(f"Confidence: {result['confidence']}")
-print(f"All probabilities: {result['all_probabilities']}")
-```
-
-## 🔍 JavaScript/TypeScript Example
-
-```javascript
-const formData = new FormData();
-formData.append('file', audioFile);
-
-const response = await fetch('http://your-space-url/predict', {
-  method: 'POST',
-  body: formData
-});
-
-const result = await response.json();
-console.log('Emotion:', result.predicted_emotion);
-console.log('Confidence:', result.confidence);
-```
-
-## ⚙️ Configuration
-
-You can modify the following in `app.py`:
-
-- **EMOTION_LABELS**: Update emotion label mappings
-- **max_duration**: Change audio duration limit (default: 3 seconds)
-- **Batch size limit**: Modify maximum files per batch request
-
-## 📊 Performance
-
-- **Inference Time**: ~100-300ms per audio file (CPU)
-- **Inference Time**: ~50-100ms per audio file (GPU)
-- **Supported Audio Length**: Up to 3 seconds (configurable)
-- **Concurrent Requests**: Supports multiple simultaneous requests
-
-## 🛠️ Troubleshooting
-
-### Common Issues
-
-1. **Model not loading**
-   - Ensure all model files are in the `model/` directory
-   - Check that file paths in `app.py` match your structure
-
-2. **Audio processing errors**
-   - Verify audio file format is supported
-   - Check that librosa and soundfile are installed correctly
-
-3. **Out of memory**
-   - Reduce batch size
-   - Use smaller audio files
-   - Enable CPU-only mode if GPU memory is limited
-
-## 📝 License
-
-This project is licensed under the MIT License.
-
-## 🙏 Acknowledgments
-
-- HuBERT model by Facebook AI Research
-- Transformers library by Hugging Face
-- FastAPI framework
-
-## 📧 Contact
-
-For questions or issues, please open an issue on GitHub or contact [your-email].
-
----
-
-**Note**: Make sure to replace `your-space-url`, `your-username`, and other placeholders with your actual information.