admin-valentin/WhisperLiveKit
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Commit graph

  • e910873312 Merge branch 'main' of https://github.com/QuentinFuxa/whisper_streaming_web Quentin Fuxa 2025-03-01 15:53:43 +0100
  • 2a869cd509 solve https://github.com/QuentinFuxa/whisper_streaming_web/issues/60#issuecomment-2692191781 Quentin Fuxa 2025-03-01 15:53:35 +0100
  • d053bac871
         huggingface steps added. Solve https://github.com/QuentinFuxa/whisper_streaming_web/issues/59 & https://github.com/QuentinFuxa/whisper_streaming_web/issues/60 Quentin Fuxa 2025-03-01 11:52:09 +0100
  • e486ef8d98
         Delete src directory Quentin Fuxa 2025-02-28 19:09:15 +0100
  • 0a1fb08371 src/web to web Quentin Fuxa 2025-02-28 18:50:00 +0100
  • ddb8860528 move files Quentin Fuxa 2025-02-28 18:49:45 +0100
  • 2e19516b3e split backends and online asr files Quentin Fuxa 2025-02-28 18:49:31 +0100
  • 3c7bc6f472 add coming soon Quentin Fuxa 2025-02-28 18:48:48 +0100
  • 2d2a4967e6 update import paths Quentin Fuxa 2025-02-28 18:41:12 +0100
  • 7e880e039e time objects is now used by DiartDiarization class Quentin Fuxa 2025-02-28 18:40:42 +0100
  • 627386a8a4 silero vad Quentin Fuxa 2025-02-28 18:39:46 +0100
  • 14af47e84b undiarized text is assigned to last speaker, with buffer information; traceback is used to format_exc errors Quentin Fuxa 2025-02-28 18:11:36 +0100
  • 00eb4a0a4f UI updated when diarization is disactivated Quentin Fuxa 2025-02-28 18:10:00 +0100
  • 2f87e592e0 Merge branch 'main' of https://github.com/QuentinFuxa/whisper_streaming_web Quentin Fuxa 2025-02-28 15:44:19 +0100
  • 56717b094f DiartDiarization now uses SpeakerSegment Quentin Fuxa 2025-02-28 15:44:09 +0100
  • 7b1c88589e // execution for diarization and transcription Quentin Fuxa 2025-02-28 15:43:46 +0100
  • 72ce8d0e3f update frontend for new variables Quentin Fuxa 2025-02-28 15:40:56 +0100
  • 09090aa3f5
         Add files via upload Quentin Fuxa 2025-02-28 11:49:32 +0100
  • d3960ffef9
         Add files via upload Quentin Fuxa 2025-02-28 11:44:29 +0100
  • 247582fb33 SpeakerSegment class for future diart improvements Quentin Fuxa 2025-02-26 21:28:22 +0100
  • 091d5d7bf5 new buffer format Quentin Fuxa 2025-02-26 21:27:39 +0100
  • 9d5d6d8031 UI improvements Quentin Fuxa 2025-02-26 21:26:24 +0100
  • 8aa3c760c7
         Add files via upload Quentin Fuxa 2025-02-24 09:27:47 +0100
  • f925ef3786
         Merge pull request #57 from QuentinFuxa/diart_integration_improvements Quentin Fuxa 2025-02-24 00:38:30 +0100
  • 2ced4fef20 diarization now works at word - not chunk - level! Quentin Fuxa 2025-02-24 00:35:42 +0100
  • 5b9b9328e0
         Merge pull request #56 from QuentinFuxa/diart_integration_improvements Quentin Fuxa 2025-02-23 23:42:37 +0100
  • d89622b9c2 Improve sentence tokenization handling - MosesSentenceSplitter now works with list input Quentin Fuxa 2025-02-23 23:41:15 +0100
  • d4096e7e11
         Merge pull request #55 from QuentinFuxa/diart_integration_improvements Quentin Fuxa 2025-02-23 23:16:10 +0100
  • 296327071d Enhance diarization logic to improve speaker attribution : corrects several bugs Quentin Fuxa 2025-02-23 23:14:10 +0100
  • 34b707d84e format html and change mapping id <-> speakers Quentin Fuxa 2025-02-23 23:13:18 +0100
  • f200f2cad4 clean diart audiosource class Quentin Fuxa 2025-02-23 23:12:40 +0100
  • 8c6d39162f Update README.md Quentin Fuxa 2025-02-20 11:20:58 +0100
  • e3adc379ed Update README.md Quentin Fuxa 2025-02-20 11:17:50 +0100
  • 90f24ef537 Update README.md Quentin Fuxa 2025-02-20 11:17:30 +0100
  • e4c84346c9 Update README.md Quentin Fuxa 2025-02-20 11:13:54 +0100
  • cf7944f13d Update README.md Quentin Fuxa 2025-02-20 10:30:47 +0100
  • d7c945dcce Merge pull request #53 from QuentinFuxa/diart_integration_improvements Quentin Fuxa 2025-02-19 16:49:21 +0100
  • fa39eda923 DiartDiarization loaded in asynccontextmanager lifespan Quentin Fuxa 2025-02-19 16:47:34 +0100
  • 01f02b066a Update README.md Quentin Fuxa 2025-02-19 15:49:30 +0100
  • a93bae69a5 Update README.md Quentin Fuxa 2025-02-19 15:39:43 +0100
  • f21dad559d Update README.md Quentin Fuxa 2025-02-19 15:28:59 +0100
  • 97c0ae6154 Merge pull request #52 from QuentinFuxa/diart_integration_improvements Quentin Fuxa 2025-02-19 14:43:39 +0100
  • 09d40a7de8 silence & diarization indicators Quentin Fuxa 2025-02-19 14:42:46 +0100
  • 2608abf0f3 Improve speaker handling; update sleep duration and manage speaker transitions more effectively Quentin Fuxa 2025-02-19 14:41:37 +0100
  • 58eba2a1f6 Enhance live transcription UI with improved speaker and silence indicators; add time information display Quentin Fuxa 2025-02-19 14:41:19 +0100
  • 450c93fef8 Merge pull request #51 from QuentinFuxa/diart_integration_improvements Quentin Fuxa 2025-02-19 11:38:59 +0100
  • 1ffa2fa224 add time duration ; display when no speaker is detected Quentin Fuxa 2025-02-19 11:26:30 +0100
  • dc24366580 Number of speakers not anymore limited to 10; a speaker has been created for "being processed" (-1), and another one for no" speaker detected" (-2) Quentin Fuxa 2025-02-19 11:25:59 +0100
  • 6121083549 add parameter to disable transcription (only diarization), add time in output Quentin Fuxa 2025-02-19 11:21:40 +0100
  • 0ecac75455 Merge pull request #47 from QuentinFuxa/logging-and-MAX_BYTES_PER_SEC Quentin Fuxa 2025-02-15 15:22:52 +0100
  • 525abcbca7 Implement logging for WebSocket events and FFmpeg process management; remove obsolete test file Quentin Fuxa 2025-02-15 15:21:33 +0100
  • 365e7c882f Merge pull request #46 from QuentinFuxa/solving-ffmpeg-process-freezing-unexpectedly Quentin Fuxa 2025-02-14 18:22:35 +0100
  • 84b09bb2cc handle ffmpeg timeouts > 5s Quentin Fuxa 2025-02-14 17:22:03 +0000
  • 4601e97221 Merge pull request #45 from QuentinFuxa/solving-ffmpeg-process-freezing-unexpectedly Quentin Fuxa 2025-02-13 00:38:45 +0100
  • 15089c80fd Add get_buffer method to retrieve unvalidated buffer in string format Quentin Fuxa 2025-02-12 23:36:52 +0000
  • 788fe1c676 optimize ffmpeg buffer reading :round duration to nearest lower 0.1s Quentin Fuxa 2025-02-12 23:28:58 +0000
  • d623578d95 Merge pull request #43 from QuentinFuxa/load-the-model-just-once Quentin Fuxa 2025-02-12 05:54:38 +0100
  • 149d2ee44c Use lifespan to load the model just one Quentin Fuxa 2025-02-12 05:53:55 +0100
  • adaca751ce remove logs handler Quentin Fuxa 2025-02-12 05:46:56 +0100
  • eb989038bd remove useless class that has been replaced by timed_objects Quentin Fuxa 2025-02-07 23:12:54 +0100
  • 1f6119e405 all text-related classes now share a common TimedText base class Quentin Fuxa 2025-02-07 23:12:04 +0100
  • f7f1f259c1 buffer is now transcript.text Quentin Fuxa 2025-02-07 12:25:05 +0100
  • b82cc3b613 adapt backend for the new classes Quentin Fuxa 2025-02-07 12:24:37 +0100
  • 46f7f9cbd1 Use Sentence, Transcript and ASRToken classes for clarity Quentin Fuxa 2025-02-07 12:24:11 +0100
  • 48c111f494 revert changes for segments buffer_trimming_way to work Quentin Fuxa 2025-02-07 10:17:45 +0100
  • 54628274d6 show language used Quentin Fuxa 2025-02-07 10:16:46 +0100
  • 0d874fb515 cuda or cpu auto detection Quentin Fuxa 2025-02-07 10:16:03 +0100
  • 4d1aa4421a Merge pull request #30 from SilasK/tsw Quentin Fuxa 2025-01-31 22:54:58 +0100
  • f4d98e2c8c Merge pull request #27 from SilasK/fix-sentencesegmenter Quentin Fuxa 2025-01-31 22:54:33 +0100
  • 15205f31d1 add doctest Silas Kieser 2025-01-28 23:17:21 +0100
  • b1f7034577 my version of timestamped text Silas Kieser 2025-01-28 23:13:15 +0100
  • 23dee02d56 sentence overflow works Silas Kieser 2025-01-28 22:38:55 +0100
  • efd80095a7 segment also works Silas Kieser 2025-01-28 22:11:28 +0100
  • f4d3df3d87 change log format Silas Kieser 2025-01-28 21:25:17 +0100
  • 9c7d429e15 add logging config to server Silas Kieser 2025-01-28 17:38:13 +0100
  • 611d33cba5 keep a test script in base directory Silas Kieser 2025-01-28 17:13:03 +0100
  • ab7c22d3e3 whisper_online works with the new sentence segment Silas Kieser 2025-01-28 17:02:21 +0100
  • 870a779666 sentence work again! Silas Kieser 2025-01-28 12:00:34 +0100
  • c3d72cae7c Merge pull request #26 from SilasK/fix-sentencesegmenter Quentin Fuxa 2025-01-28 15:53:26 +0100
  • 4622fe7aff Merge branch 'main' into fix-sentencesegmenter Quentin Fuxa 2025-01-28 15:53:10 +0100
  • 8ee1488c08 rename to_flush to concatenate_tsw Silas Kieser 2025-01-27 16:49:22 +0100
  • 77d43885a3 chunk at sentence takes now an argument =self.comited Silas Kieser 2025-01-27 16:29:06 +0100
  • 04170153e0 improve logging Silas Kieser 2025-01-27 16:12:30 +0100
  • baddf0284b buffer length in sentence segmentation is no also max as in segment. Silas Kieser 2025-01-27 15:36:19 +0100
  • 6e0f1dda25 Merge remote-tracking branch 'contrib/fix-sentencesegmenter' Quentin Fuxa 2025-01-26 15:34:41 +0100
  • c66794e1f5 Merge pull request #20 from SilasK/clean-main Quentin Fuxa 2025-01-26 14:57:52 +0100
  • f0eaffacd3 improve logging in whisper_online.py Silas Kieser 2025-01-21 14:59:36 +0100
  • 69a2ed6bfb add logger for online asr Silas Kieser 2025-01-21 14:45:45 +0100
  • 25eb276794 ignore wav and scripts Silas Kieser 2025-01-21 14:08:41 +0100
  • 9f262813ec sep for mlx is also "" Silas Kieser 2025-01-21 10:39:38 +0100
  • 4293580581 use moses sentence segmenter instead of tokenizer Silas Kieser 2025-01-21 12:06:03 +0100
  • 42d2784c20 clearer log messages for sentence segmentation Silas Kieser 2025-01-21 11:18:54 +0100
  • 7fad0a3ee2 sep for mlx is also "" Silas Kieser 2025-01-21 10:39:38 +0100
  • 27d2db77f7 Update README.md Quentin Fuxa 2025-01-20 03:08:01 +0100
  • fba37eba0a move to src Quentin Fuxa 2025-01-19 21:17:55 +0100
  • 5523b51fd7 first speaker is "0" no more None Quentin Fuxa 2025-01-19 19:40:09 +0100
  • 9bdb92e923 update demo.png Quentin Fuxa 2025-01-19 19:36:10 +0100
  • b51c8427f4 diart link added Quentin Fuxa 2025-01-19 17:12:55 +0100
  • 977436622a add diarization (beta). Disabled by default Quentin Fuxa 2025-01-19 17:12:40 +0100
  • ce56264241 split whisper_online.py into smaller files Quentin Fuxa 2025-01-14 20:52:53 +0100