ysharma3501/LavaSR
Restores speech at 4000x realtime on GPU
View on index · View in 3D Map
Related repos
- Trains text-to-speech models with 2000x realtime generation on GPU
- Real-time transcription using Faster-Whisper
- Speaker diarization for audio streams in real time
- Voice chat using WebGPU for local inference
- Production-ready text-to-speech API with multiple voices and a real-time dashboard
- Video LLM for real-time commentary with streaming speech transcription
- LLM-based text-to-speech training and voice cloning with multi-GPU support
- Clones voices at 150x realtime speed
- Voice activity and audio event detection for 100+ languages
- AI voice cloning and text-to-speech with quick voice samples
