Skip to content

v0.105.2

Choose a tag to compare

View all tags
@basnijholt basnijholt released this 15 Sep 07:56
v0.105.2
1128990
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Bug fixes

  • Accept M4A, OGG, MP3, and other FFmpeg-supported uploads in the Transformers ASR backend. Transformers and NeMo now share audio preparation, with invalid input reported as HTTP 400. (#644)
  • Normalize stereo, non-16-bit, and non-16 kHz WAV uploads before WAV-only ASR backends read them, preventing incorrect sample interpretation. Standard 16 kHz mono 16-bit WAV uploads pass through unchanged. (#644)
  • Stop FFmpeg conversions when requests are cancelled or conversion times out, then clean up temporary files. Faster-whisper retains native compressed-audio decoding without a system FFmpeg dependency. (#644)

Full changelog: v0.105.1...v0.105.2

Assets 4
Loading