v1.0.2

What's Changed

• Set debug = false by default by @gaspardpetit in #51
• Do not use ami diarization by default by @gaspardpetit in #52
• Simplify configs by @gaspardpetit in #53
• Allow reusing models across multiple transcriptions to avoid loading-… by @gaspardpetit in #54
• Configure whisper transcription from function parameters by @gaspardpetit in #55
• Improve vad filtering by @gaspardpetit in #57
• Invalidate prefix when language changes by @gaspardpetit in #58
• Improve language detection with fallback by @gaspardpetit in #59

Full Changelog: v1.0.1...v1.0.2

1.0.1

What's Changed

• Streaming approach to transcription by @gaspardpetit in #41
• fix: support for specific number of speakers when performing diarization by @gaspardpetit in #43
• Improve export format handling and output directory support by @gaspardpetit in #42
• Print unacknowledged utterances by @gaspardpetit in #44
• Commit unacknowledged words and utterances when discarding them from … by @gaspardpetit in #45
• Fix detect number of speakers by @gaspardpetit in #46
• Use tuple in ass export by @gaspardpetit in #47
• fix: support start and to arguments with minutes and seconds only by @gaspardpetit in #49
• Stream from ffmpeg by @gaspardpetit in #48
• improve handling when diarization is misson or no words have been gen… by @gaspardpetit in #50

Full Changelog: v0.1.6...v1.0.1

0.1.6

What's Changed

• Fixed pip install -r requirements.txt by @gaspardpetit in #31
• Make sure device preference is forwarded to all the calls by @gaspardpetit in #34
• ensure that all temp files are under the working directory by @gaspardpetit in #32
• When pyannote model cannot be downloaded, skip diarization by @gaspardpetit in #33
• When ffmpeg is not found, fallback on SoundAudio by @gaspardpetit in #35
• Added transcribe_only flag to skip prefiltering and diarization by @gaspardpetit in #36
• Added support for configuring the number of beams using -b or --nb_beams by @gaspardpetit in #38
• Set version to 0.1.6 by @gaspardpetit in #37

Full Changelog: v0.1.5...v0.1.6

0.1.5

What's Changed

• Use faster whisper for transcription by default by @gaspardpetit in #21
• Version changed to 0.1.4 by @gaspardpetit in #20
• Attempt to improve prompt to avoid capitalization issues by @gaspardpetit in #22
• increase beam size to 9 for better results by @gaspardpetit in #23
• Disable the torchaudio warnings in pyannote and torch_audiomentations by @gaspardpetit in #24
• Document the gating of the pyannote models on Hugging Face by @gaspardpetit in #25
• Expose the models to settings so they can be changed by @gaspardpetit in #26
• Standardize the execution of pipeline filters by @gaspardpetit in #27
• Skip language detection when only one language is selected by @gaspardpetit in #28
• Allow loading and unloading of models by @gaspardpetit in #30
• Updated version to 0.1.5 by @gaspardpetit in #29

Full Changelog: v0.1.3...v0.1.5

0.1.3

What's Changed

• Avoid exception when attempting to perform negative padding on long a… by @gaspardpetit in #15
• Updated dockerfile to properly install requirements by @gaspardpetit in #16
• Write to stdout by @gaspardpetit in #17
• Add dockerfile by @gaspardpetit in #18
• Updated version to 0.1.3 by @gaspardpetit in #19

Full Changelog: v0.1.2...v0.1.3

0.1.2

Full Changelog: v0.1.1...v0.1.2

0.1.1

What's Changed

• Adding Dockerfile, build and run scripts by @gaspardpetit in #10
• Do not use hardcoded "out" directory as the working directory by @gaspardpetit in #12
• Ensure logs go to stderr and output results to stdout by @gaspardpetit in #13
• Updated version, filter requirements by @gaspardpetit in #14

Full Changelog: v0.1.0...v0.1.1

0.1.0

What's Changed

• Work around issue speechbrain/speechbrain#2334 by @gaspardpetit in #3
• Cleanup the use of kwargs arguments by @gaspardpetit in #4
• Implement a main() entrypoint by @gaspardpetit in #6
• Added setting and support for running the pipeline entirely on CPU by @gaspardpetit in #7
• Added setting and support for running the pipeline entirely on CPU by @gaspardpetit in #8
• Added command line tool and usage documentation by @gaspardpetit in #9

Full Changelog: v0.0.1...v0.1.0

0.0.1

What's Changed

• Build fix for linter by @gaspardpetit in #1
• Cicd fix build by @gaspardpetit in #2

New Contributors

• @gaspardpetit made their first contribution in #1

Full Changelog: https://github.com/gaspardpetit/verbatim/commits/v0.0.1