v0.5.7

Fixed

• fix: add console script entry point

Full Changelog: v0.5.6...v0.5.7

v0.5.6

• Allows passing VAD models to device passed to constructor

v0.5.5

• fix: allow kwarg passing to segment methods

v0.5.4

fix: keep huggingfacehub imports local

Full Changelog: v0.5.3...v0.5.4

v0.5.3

What's Changed

• refactor[torchcodec]: move torchcodec to optional by @winsomeglint in #25

Full Changelog: v0.4.2...v0.5.3

v0.5.2

• Fix pyannote-audio vad; now uses segmentation-3.0 model

v0.5.1

Full Changelog: v0.5.0...v0.5.1

• fix: add huggingfacehub as a base dependency

v0.5.0

What's Changed

• ⚠️ feature: add a caul model cache <model_name> <cache-dir> CLI endpoint by @ClemDoum in #22
• refactor: get rid of torchaudio by @ClemDoum in #23

Full Changelog: v0.4.1...v0.5.0

v0.4.1

Fixed

• non local torch import in utils

Full Changelog: v0.4.0...v0.4.1

0.4.0

• Adds FireRedAsr2s and faster-whisper pipelines
• Chunks long audio files (default is 30mb) rather than loading them into memory