Skip to content

41% smaller encoder than Whisper Large v3 Turbo via LiteASR low-rank approximation. 99 languages. Fills the gap between Whisper Small and Large v3 Turbo.

WebGPU
WASM
Streaming
Apache-2.0

Try It

LiteWhisper Large v3 Turbo
Not loaded(400.0 MB)

Download the model first to start recording

Technical Details

Architecture

Whisper (LiteASR low-rank)

Parameters

600M

Model Size

400 MB

Languages

en, fr, de, es, it, ja, ko, zh, pt, ru

Links