RealFeedApp beta
- Whistle recognizes speech on mobiles, wearables, and robots
The model transcribes audio, provides word timestamps, and generates speech embeddings, operating within a single 16.9 MB file on the CPU.
- Field
- tools and releases
- What they did
- Released the Whistle speech recognition model, a 16.9 MB file that runs on mobile devices and robots without external dependencies. It performs transcription, word timestamps, and speech embedding directly on the device.
- Why it matters
- This enables resource-constrained hardware like microcontrollers and wearables to process audio locally, avoiding the need to send data to the cloud.
#speech recognition#llm#on-device#whistle
