StreamingSpeakerDiarization Versions Save

New Features!

Major changes in this new version! Including compatibility with pyannote 3.*, SpeechBrain, WeSpeaker and NeMo embedding models, totaling 8 new models to create speaker diarization pipelines and 1 new model for voice activity detection.

This version also adds compatibility with ONNX models and a documentation page at diart.readthedocs.io

What's Changed

• Add compatibility with pyannote 3.0 embedding wrappers by @sorgfresser in #188
• Add support for powerset segmentation model by @hbredin in #198
• Add ONNX compatibility by @juanmc2005 in #204
• README improvements by @juanmc2005 in #207
• Add documentation page by @juanmc2005 in #209
• Make ONNX runtime optional by @juanmc2005 in #215
• Add reproducibility warning in README by @juanmc2005 in #216

New contributors

Thank you @sorgfresser for your huge contribution in #188 !

Full Changelog: https://github.com/juanmc2005/diart/compare/v0.8...v0.9

What's Changed

• Voice Activity Detection by @juanmc2005 in #143
• Fix sample rate issues by @juanmc2005 in #153
• [joss] Add Conda environment by @sneakers-the-rat in #172
• [joss] Use ABC and abstractmethod rather than NotImplementedError by @sneakers-the-rat @juanmc2005 in #173 #179
• Fix torchaudio version incompatibility by @juanmc2005 in #181
• Add special audio source for Apple devices by @juanmc2005 in #182
• Catch keyboard interrupt in diart.stream by @juanmc2005 in #183
• Remove PipelineConfig.from_dict() by @juanmc2005 in #189
• Fix link to reproducibility section by @hbredin in #191

New Contributors

Thank you @sneakers-the-rat for your extremely valuable feedback and help as part of the JOSS review!

Full Changelog: https://github.com/juanmc2005/diart/compare/v0.7...v0.8

What's Changed

• Plug & play speaker diarization pipelines by @juanmc2005 in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/101
• Parallel execution of Benchmark by @juanmc2005 in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/124
• Better installation instructions by @juanmc2005 in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/130
• Rewrite websockets by @juanmc2005 in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/131
• Add left padding to short audio files by @juanmc2005 in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/135
• Allow to select microphone device by @juanmc2005 in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/136
• Update logo and demo gif by @juanmc2005 in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/137
• Shift pipeline timestamps by @juanmc2005 in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/138

Full Changelog: https://github.com/juanmc2005/StreamingSpeakerDiarization/compare/v0.6...v0.7

What's Changed

• Compatibility with torchaudio streams by @juanmc2005 in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/91
• Online speaker diarization as a block by @juanmc2005 in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/92
• Fix bug: RTTM output not being patched when closing plot window by @juanmc2005 in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/100
• Add cropping_mode to DelayedAggregation by @bhigy in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/105
• Compatibility with pyannote.audio 2.1.1 requirements by @juanmc2005 in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/108

New Contributors

• Thank you @bhigy for the bug hunting in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/105!

Full Changelog: https://github.com/juanmc2005/StreamingSpeakerDiarization/compare/v0.5.1...v0.6

What's Changed

• Fix wrong config reference and unpatched annotation by @juanmc2005 in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/89

Full Changelog: https://github.com/juanmc2005/StreamingSpeakerDiarization/compare/v0.5...v0.5.1

What's Changed

• Add study_or_path as a Path for conversion from string by @AMITKESARI2000 in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/74
• Update WebSocketAudioSource by @ckliao-nccu in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/78
• Fix bug with empty RTTMs by @zaouk in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/81
• Add websocket compatibility + other improvements by @juanmc2005 in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/77
• Export csv report in diart.benchmark when output is provided by @juanmc2005 in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/86

New Contributors

• @AMITKESARI2000 made their first contribution in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/74
• @ckliao-nccu made their first contribution in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/78

Acknowledgements

Thank you @AMITKESARI2000, @ckliao-nccu and @zaouk for all the bug fixes!

Full Changelog: https://github.com/juanmc2005/StreamingSpeakerDiarization/compare/v0.4...v0.5

What's Changed

• Replace resolve_features with TemporalFeatureFormatter by @juanmc2005 in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/59
• Make pyannote.audio optional (still mandatory to run default pipeline) by @juanmc2005 in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/61
• Minor features and improvements by @juanmc2005 in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/64
• Adds documentation for some of the classes and methods by @zaouk in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/31
• Add hyper-parameter tuning with optuna by @juanmc2005 in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/65

New Contributors

• Thank you @zaouk for your contribution in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/31 !

Full Changelog: https://github.com/juanmc2005/StreamingSpeakerDiarization/compare/v0.3...v0.4

What's Changed

• Python 3.7 compatibility and PortAudio error fix by @Yagna24 in #29
• Add citation by @hbredin in #38
• Benchmark script + improvements and bug fixes by @juanmc2005 in #46
• Improve API names by @juanmc2005 in #47
• Add OverlapAwareSpeakerEmbedding class by @juanmc2005 in #51
• Add inference API with RealTimeInference and Benchmark by @juanmc2005 in #55

New Contributors

• Thank you @Yagna24 for your contribution in python 3.7 compatibility!

Full Changelog: https://github.com/juanmc2005/StreamingSpeakerDiarization/compare/v0.2.1...v0.3

What's Changed

• Fix empty segment in buffer_output causing a crash by @juanmc2005 in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/24

Full Changelog: https://github.com/juanmc2005/StreamingSpeakerDiarization/compare/v0.2...v0.2.1

What's Changed

• Replace operators.aggregate() with functional.DelayedAggregation by @juanmc2005 in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/16
• Add Hamming-weighted average to DelayedAggregation by @juanmc2005 in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/18
• Asynchronous microphone reading by @juanmc2005 in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/19
• Modular OutputBuilder + better demo performance by @juanmc2005 in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/20
• Improve README by @juanmc2005 in https://github.com/juanmc2005/StreamingSpeakerDiarization/pull/21

Full Changelog: https://github.com/juanmc2005/StreamingSpeakerDiarization/compare/v0.1...v0.2