shashikg/WhisperS2T
Optimizes the Whisper speech-to-text model with multiple inference backends including TensorRT for faster transcription.
Not currently ranked — collecting fresh signals.
star history
WhisperS2T is an optimized speech-to-text pipeline built around OpenAI’s Whisper model. It supports multiple inference engines such as TensorRT and TensorRT-LLM, and includes voice activity detection to improve transcription speed and accuracy. The project claims 2.3X speed improvement over WhisperX and 3X over HuggingFace’s implementation with FlashAttention 2.
Frequently asked
- What is shashikg/WhisperS2T?
- Optimizes the Whisper speech-to-text model with multiple inference backends including TensorRT for faster transcription.
- Is WhisperS2T open source?
- Yes — shashikg/WhisperS2T is open source, released under the MIT license.
- What language is WhisperS2T written in?
- shashikg/WhisperS2T is primarily written in Jupyter Notebook.
- How popular is WhisperS2T?
- shashikg/WhisperS2T has 577 stars on GitHub.
- Where can I find WhisperS2T?
- shashikg/WhisperS2T is on GitHub at https://github.com/shashikg/WhisperS2T.