sanchit-gandhi

3 ms·

1. 1.▲
   Fine-tune Wav2Vec2-BERT for low resource speech recognition(huggingface.co)

   2 points by sanchit-gandhi 3y ago|1 comments
2. 2.▲
   by sanchit-gandhi 3y ago

   A month ago, Meta AI released Wav2Vec-Bert, one of the building blocks of the powerful Seamless Communications models. The checkpoint is MIT licensed and available in Hugging Face Transformers, where you can fine-tune it to get comparable s
3. 3.▲
   by sanchit-gandhi 3y ago

   Hugging Face Whisper (the backend to insanely-fast-whisper) now supports PyTorch SDPA attention with PyTorch>=2.1.1 It's enabled by default with the latest Transformers version, so just make sure you have: * torch>=2.1.1 * transf
4. 4.▲
   by sanchit-gandhi 3y ago

   Indeed, insanely-fast-whisper supports beam-search with a small code modification to this code snippet: https://huggingface.co/openai/whisper-large-v3 Just call the pipeline with: result = pipe(sample, generate_kwargs
5. 5.▲
   by sanchit-gandhi 3y ago

   There's no need to wait for MusicGen to generate the full audio before you can start listening to the outputs With streaming, you can play the audio as soon as the first chunk is ready In practice, this reduces the latency to just 5s
6. 6.▲
   Faster MusicGen Generation with Streaming(huggingface.co)

   3 points by sanchit-gandhi 3y ago|1 comments