v0.2.24

mikeqzy/3dgs-avatar-releasev0.2.24Jul 11, 2026by QuentinFuxa

AI Summary

Adds a new AlignAtt translation backend for streaming LLM translation and updates benchmarks. Includes fixes for MLX-Whisper empty captions and improves error handling to fail loudly when ASR produces no output.

Key Highlights

  • New AlignAtt translation backend for streaming LLM translation.
  • Benchmark figures re-measured on H100.
  • Fixes for MLX-Whisper empty captions and loud failure handling.

New Features

  • AlignAtt translation backend
  • Benchmark figures re-measurement
  • MLX-Whisper fixes
  • Loud failure handling for ASR

Full Release Notes

Streaming LLM translation, refreshed benchmarks and loud failures.

## Added
- AlignAtt translation backend (`--translation-backend alignatt`): streaming LLM translation through an [Alignatt4LLM](https://github.com/QuentinFuxa/Alignatt4LLM) sidecar. The model drafts ahead over the unstable ASR tail and commits only target words whose attention lands on committed source words, so translations are append-only and released the instant the ASR commits. See [docs/translation-alignatt.md](https://github.com/QuentinFuxa/WhisperLiveKit/blob/main/docs/translation-alignatt.md).
- Benchmark figures re-measured on an H100 with the current backends (the README scatter plots dated from March and misrepresented the qwen3 backends). Raw results in `benchmarks/h100_scatter/`.

## Fixed
- mlx-whisper produced empty captions with torch >= 2.13 (#383, thanks @RobertBartelds-FleetEnergies).
- The server now fails loudly when the ASR backend produces nothing: warmup errors abort startup with the cause, and a watchdog logs an explicit error if audio flows but no text is ever produced.