https://github.com/alumae/kaldi-offline-transcriber

sort by:
Revision Author Date Message Commit Date
0aec0b9 added --do-language-id option 04 August 2022, 08:39:43 UTC
928adff Skip very short segments 29 June 2022, 09:43:19 UTC
adfa8de Merge branch 'master' of github.com:alumae/kaldi-offline-transcriber 27 May 2022, 12:06:00 UTC
0ab0929 Avoid 0-length segments 27 May 2022, 12:05:49 UTC
9e0142b Update README.md 26 May 2022, 06:21:46 UTC
8815439 Update README.md 26 May 2022, 06:21:16 UTC
9028b47 Removed SAD 25 May 2022, 11:39:16 UTC
1c2068f Now uses Silero VAD for speech activity detection 24 May 2022, 13:19:08 UTC
51bfdb3 Lates updates 15 June 2021, 10:50:12 UTC
c88870b Now works with recent changes 14 June 2021, 07:10:31 UTC
dbe5ad4 Some fixes 11 June 2021, 21:15:11 UTC
7dee822 Updating to newer versions, adding language ID 11 June 2021, 14:29:23 UTC
2ddfbb0 Bug fix 04 June 2021, 15:54:41 UTC
caadc22 Docker image now inherits from kaldi's Docker image 22 February 2021, 13:30:07 UTC
d3b8372 Update README.md Update name of the models file 28 January 2021, 12:47:05 UTC
14398f8 small fixes to speaker ID server stuff 05 June 2019, 15:28:03 UTC
9b74727 Merge branch 'master' of /home/tanel/devel/kaldi-offline-transcriber 05 June 2019, 13:39:43 UTC
82036ff reverted some changes 05 June 2019, 13:39:39 UTC
2059fe1 Small changes related to words-to-numbers 05 June 2019, 13:38:37 UTC
918806d Merge branch 'master' of /home/tanel/devel/kaldi-offline-transcriber 05 June 2019, 13:28:58 UTC
19dea83 Can now use external speaker ID server 05 June 2019, 13:19:47 UTC
e2dbf72 Riigikogu spetsiifilised muudatused 05 June 2019, 06:50:56 UTC
d64a8e7 Added Sync elements after each Turn beginning to make TSAB happy 08 January 2019, 14:11:05 UTC
dc16cd0 Added words-to-numbers conversion to appropriate places 29 November 2018, 11:06:58 UTC
a4b93d3 Program that converts words to numbers, using Pynini 27 November 2018, 10:20:20 UTC
6e84dbf Program that postprocesses JSON-formatted transcript using an external program 27 November 2018, 10:19:37 UTC
79f0d38 added confidence scores to CTM and JSON result files 22 November 2018, 15:33:13 UTC
b980ff6 script for converting words to numbers using FST 22 November 2018, 15:32:03 UTC
ad6ff46 Avoid small gaps between turns 14 November 2018, 14:43:41 UTC
b1d10b8 Merge branch 'master' of github.com:alumae/kaldi-offline-transcriber 14 November 2018, 14:11:35 UTC
b03a80a Fixed a Transcriber file format issue 14 November 2018, 14:10:58 UTC
142f0aa Update README.md 01 November 2018, 15:14:47 UTC
97152d5 minor fixes 31 October 2018, 16:25:30 UTC
b67c07b Some minor fixes 31 October 2018, 14:59:35 UTC
b656610 fix typo 31 October 2018, 14:58:01 UTC
f7dad1a Refactored and introduced a new JSON format that holds all information about word and segment timings 31 October 2018, 11:28:12 UTC
fe0ab0f Refactored and introduced a new JSON format that holds all information about word and segment timings 31 October 2018, 11:04:16 UTC
623d5c9 Update README.md 28 October 2018, 12:47:57 UTC
bcec748 Merge branch 'master' of github.com:alumae/kaldi-offline-transcriber 21 October 2018, 18:27:20 UTC
df8d5e8 much faster compounding now 21 October 2018, 18:26:43 UTC
39a19df clean up temp file in src-audio 21 October 2018, 18:26:21 UTC
4962ba9 Update README.md 09 October 2018, 12:00:31 UTC
4b02ef5 Update README.md Changed public image URL. 09 October 2018, 11:55:44 UTC
80256af delete wav from src-audio when cleaning up 04 October 2018, 07:15:32 UTC
c2c99b9 Updated SID models 12 September 2018, 18:30:23 UTC
19541b7 Added Dockerfile for pre-building Estonian system 31 August 2018, 08:58:59 UTC
09bc2a8 Added Dockerfile for pre-building Estonian system 31 August 2018, 08:54:49 UTC
9be252a Added Dockerfile for pre-building Estonian system 31 August 2018, 08:23:57 UTC
61b16fd Added Dockerfile for pre-building Estonian system 31 August 2018, 08:22:32 UTC
a532032 cleanup after creating decoding graph 24 August 2018, 13:08:43 UTC
7d511f1 cleanup after init 24 August 2018, 09:04:38 UTC
64b71d0 Some minor improvements 23 August 2018, 05:36:34 UTC
b2a457d Speaker ID system now uses Kaldi's native i-vector scoring 21 August 2018, 13:52:45 UTC
c8fb6fe Titlecase after questions 09 August 2018, 12:52:10 UTC
924aca4 removed reduntant option 09 August 2018, 11:01:01 UTC
25a5116 doc updates 08 August 2018, 11:42:44 UTC
8a71397 doc updates 08 August 2018, 11:34:19 UTC
3f3f731 Updates related to segments file 08 August 2018, 10:33:23 UTC
11bfbd9 Refactored -- now segments file is used, instead of splitting the wav ühysically to pieces 08 August 2018, 08:56:53 UTC
194e124 Fixes related to init 07 August 2018, 11:20:06 UTC
d5982b1 Now uses LM with special unk handling, and <unk> words can be reconstructed from pronuciation 07 August 2018, 10:32:54 UTC
5995e98 Use ffmpeg for decoding mp4 files 28 June 2017, 08:27:39 UTC
2cc26c6 Added support for SubRip subtitle files (.srt) 15 June 2017, 06:58:07 UTC
b4395b6 Now uses DNN-based speaker ID, trained in a weakly unsupervised manner. Requires Keras 30 May 2017, 11:47:10 UTC
59f935b Now uses DNN-based speaker ID, trained in a weakly unsupervised manner. Requires Keras 29 May 2017, 13:38:30 UTC
8e5c513 mistakenly commited 29 May 2017, 13:37:00 UTC
81f77c6 Now uses DNN-based speaker ID, trained in a weakly unsupervised manner. Requires Keras 29 May 2017, 13:15:27 UTC
8050527 Replaced pyfst in compounder.py with OpenFst's native extension 02 May 2017, 12:21:39 UTC
c9cdfb6 Replaced pyfst in compounder.py with OpenFst's native extension 02 May 2017, 12:20:29 UTC
50cf17b Fixes for chain model CTM generation, too long subtitles in SBV file, and out-of-memory errors for diarization of long audio files 26 April 2017, 12:55:54 UTC
c745dce Fixes titlecasing bug 20 February 2017, 14:36:07 UTC
a36fa2d Fixes for some cases when transcribing is restarted 16 February 2017, 08:13:54 UTC
323fa1a Compatibiliy to user-set LD_LIBRARY_PATH 13 February 2017, 12:34:52 UTC
87ff3c1 Migrated to chain models, made python3 compatible 13 February 2017, 12:19:12 UTC
708bcfa Compatibiliy to user-set LD_LIBRARY_PATH 13 February 2017, 12:14:35 UTC
e12288a Migrated to chain models, made python3 compatible 13 February 2017, 11:59:12 UTC
17d5ec5 Migrated to chain models, made python3 compatible 13 February 2017, 10:21:06 UTC
066a5db Migrated to chain models, made python3 compatible 13 February 2017, 10:19:38 UTC
86e3c32 Updated models 30 December 2015, 01:06:21 UTC
b3e0552 Updated models 30 December 2015, 01:03:48 UTC
780eae1 Compability fix 29 December 2015, 22:11:24 UTC
a90f62d Cosmetic fixes 04 December 2015, 14:34:08 UTC
eb00e3b mp3 files are converted now using ffmpeg which is more robust to exotic formats 25 November 2015, 17:01:06 UTC
bb7a3b4 Fix for an issue that caused speaker ID to be executed when it was turned off 08 September 2015, 15:20:47 UTC
6730923 About memory reqs 14 May 2015, 11:28:27 UTC
2bede6f fix in .init 14 May 2015, 10:10:09 UTC
cb389ed Cosmetic fixes 14 May 2015, 10:03:14 UTC
343405f Removed option to decode with old-style nnet2 models, made decodining with online nnet2 models multithreaded 14 May 2015, 09:32:24 UTC
14f9ddf bug fix 16 March 2015, 12:59:10 UTC
48ef75c Updated LM 11 March 2015, 14:03:41 UTC
5300497 Clarified that the system is currently for Estonian 10 March 2015, 15:45:35 UTC
2feec82 small utf-8 related fix 06 March 2015, 09:01:59 UTC
babf185 titlecases words after . 05 March 2015, 10:52:18 UTC
9c4b7f2 small bug fix 03 March 2015, 18:59:42 UTC
594b21a txt files now have (optional) punctuation 03 March 2015, 13:35:28 UTC
0b3f213 small bug fix 03 March 2015, 13:25:59 UTC
722dc37 bug fix 03 March 2015, 13:10:25 UTC
56db315 Integrated punctuation insertion module 03 March 2015, 11:38:33 UTC
425267f Added optional support for automatic punctuation (NB! uses SRILM currently -- not for commercial use) 29 December 2014, 14:47:55 UTC
dc2107e Added optional support for automatic punctuation (NB! uses SRILM currently -- not for commercial use) 29 December 2014, 14:36:32 UTC
back to top