# ML Models Directory

Place the following TFLite models in this directory:

1. `gemma.tflite` - Gemma 2B model for text conversations
2. `fastspeech2.tflite` - FastSpeech 2 model for speech synthesis
3. `mobilenet_v2.tflite` - MobileNetV2 model for visual processing

## Model Specifications

### Gemma 2B
- Input: Text tokens (max length 512)
- Output: Text generation
- Size: ~2GB

### FastSpeech 2
- Input: Text sequence
- Output: Mel spectrogram for voice synthesis
- Features: Male/female voice options

### MobileNetV2
- Input: 224x224 RGB image
- Output: 1000-class classification
- Features: Optimized for mobile devices