Ecosyste.ms: Issues

An open API service for providing issue and pull request metadata for open source projects.

GitHub / openai/whisper issues and pull requests

#1244 - hola

Pull Request - State: closed - Opened by erickmauri about 1 year ago

#1242 - Drop ffmpeg-python dependency and call ffmpeg directly.

Pull Request - State: closed - Opened by petterreinholdtsen about 1 year ago - 9 comments

#1239 - Formatted Warnings Output

Pull Request - State: closed - Opened by brett-b112 about 1 year ago - 1 comment

#1236 - Updated README.md to provide more insight on BLEU and specific appendices

Pull Request - State: closed - Opened by brett-b112 about 1 year ago - 1 comment

#1233 - Fix numba depreceation notice

Pull Request - State: closed - Opened by m3at about 1 year ago

#1225 - Add new `job_details.model` key to transcribe return dict

Pull Request - State: open - Opened by ururk about 1 year ago - 2 comments

#1219 - Update decoding.py

Pull Request - State: closed - Opened by jongwook about 1 year ago

#1211 - workflows/python-publish.yml: bump actions version to fix node warning

Pull Request - State: closed - Opened by kbdharun about 1 year ago

#1196 - Update README.md about ffmpeg PATH variables

Pull Request - State: open - Opened by SFARPak about 1 year ago

#1184 - Implement max line width and max line count, and make word highlighting optional

Pull Request - State: closed - Opened by ryanheise about 1 year ago - 5 comments

#1180 - added cantonese to the language list

Pull Request - State: closed - Opened by Keith-Hon about 1 year ago - 1 comment

#1178 - Add Cantonese to the language list

Pull Request - State: closed - Opened by Keith-Hon about 1 year ago

#1171 - Python 3.11

Pull Request - State: closed - Opened by johnnynunez over 1 year ago - 5 comments

#1163 - Update tokenizer.py

Pull Request - State: closed - Opened by jongwook over 1 year ago

#1155 - Update decoding.py

Pull Request - State: closed - Opened by FernanOrtega over 1 year ago - 2 comments

#1154 - Align word to proper segment

Pull Request - State: closed - Opened by doublex over 1 year ago

#1135 - IndexError: arrays used as indices must be of integer type

Pull Request - State: closed - Opened by doublex over 1 year ago - 1 comment

#1134 - IndexError: arrays used as indices must be of integer type

Pull Request - State: closed - Opened by doublex over 1 year ago

#1133 - Create maison8

Pull Request - State: closed - Opened by ghost over 1 year ago

#1123 - docs(readme): remove instructions for installing huggingface tokenizer

Pull Request - State: closed - Opened by debloper over 1 year ago - 1 comment

#1119 - Per Token Confidence + Color terminal example

Pull Request - State: open - Opened by SinanAkkoyun over 1 year ago - 13 comments

#1114 - Squash long words at window and sentence boundaries.

Pull Request - State: closed - Opened by ryanheise over 1 year ago - 5 comments

#1105 - Update README.md to reference tiktoken

Pull Request - State: closed - Opened by bushyn over 1 year ago

#1090 - abort find_alignment on empty input

Pull Request - State: closed - Opened by jongwook over 1 year ago

#1089 - Fix truncated words list when the replacement character is decoded

Pull Request - State: closed - Opened by guillaumekln over 1 year ago - 1 comment

#1087 - Fix alignment between the segments and the list of words

Pull Request - State: closed - Opened by guillaumekln over 1 year ago - 1 comment

#1076 - Fix github language stats dominated by jupyter notebook

Pull Request - State: closed - Opened by akashmjn over 1 year ago - 2 comments

#1061 - kwargs in decode() for convenience

Pull Request - State: closed - Opened by jongwook over 1 year ago

#1060 - fix all_tokens handling

Pull Request - State: closed - Opened by jongwook over 1 year ago

#1053 - Use triton==2.0.0

Pull Request - State: closed - Opened by jongwook over 1 year ago - 6 comments

#1052 - attempt to fix the repetition/hallucination issue identified in #1046

Pull Request - State: closed - Opened by jongwook over 1 year ago - 15 comments

#1051 - Try installing triton only if linux & x86_64

Pull Request - State: closed - Opened by jongwook over 1 year ago

#1044 - Use tiktoken

Pull Request - State: closed - Opened by jongwook over 1 year ago - 1 comment

#1040 - add always_use_initial_prompt

Pull Request - State: open - Opened by mercury233 over 1 year ago - 12 comments

#1039 - support comma separated output_format

Pull Request - State: open - Opened by mercury233 over 1 year ago - 3 comments

#1038 - apply formatting with `black`, `isort`, and `flake8`

Pull Request - State: closed - Opened by jongwook over 1 year ago

#1033 - Decoding improvements

Pull Request - State: closed - Opened by jongwook over 1 year ago

#1025 - added progress_callback in transcribe method

Pull Request - State: open - Opened by jhj0517 over 1 year ago - 7 comments

#1023 - Add progress callback

Pull Request - State: closed - Opened by jhj0517 over 1 year ago

#1021 - remove auxiliary audio extension

Pull Request - State: closed - Opened by ain-soph over 1 year ago - 1 comment

#1012 - add task for transliteration to English

Pull Request - State: closed - Opened by spgoswami1 over 1 year ago - 1 comment

#1005 - Update README.md

Pull Request - State: closed - Opened by doggy8088 over 1 year ago - 1 comment

#997 - condition_on_previous was being partially ignored

Pull Request - State: closed - Opened by jmward01 over 1 year ago

#991 - Add per token confidence to each segment.

Pull Request - State: open - Opened by Pikauba over 1 year ago - 2 comments

#973 - Expose punctuation options in cli and transcribe()

Pull Request - State: closed - Opened by ryanheise over 1 year ago - 2 comments

#965 - Update model-card.md

Pull Request - State: closed - Opened by mustafakentar over 1 year ago - 2 comments

#934 - Improve the seeking algorithm

Pull Request - State: closed - Opened by jumon over 1 year ago - 4 comments

#914 - Fix infinite loop caused by incorrect timestamp tokens prediction

Pull Request - State: closed - Opened by andrewchernyh over 1 year ago - 1 comment

#907 - Reconfigure output to utf-8

Pull Request - State: open - Opened by sanxfxteam over 1 year ago

#894 - Update README.md

Pull Request - State: closed - Opened by roman-vasi1enko over 1 year ago

#889 - drop python 3.7 support

Pull Request - State: closed - Opened by jongwook over 1 year ago - 1 comment

#888 - adding word confidence score computation

Pull Request - State: closed - Opened by emirdemirel over 1 year ago

#887 - handle printing even if sys.stdout.buffer is not available

Pull Request - State: closed - Opened by jongwook over 1 year ago

#881 - Create Sasha Trailer

Pull Request - State: closed - Opened by SiobhanMcHugh over 1 year ago

#869 - word-level timestamps in `transcribe()`

Pull Request - State: closed - Opened by jongwook over 1 year ago - 15 comments

#867 - use stdout for printing transcription progress

Pull Request - State: closed - Opened by jongwook over 1 year ago

#864 - Handle XDG_CACHE_HOME properly for download_root

Pull Request - State: closed - Opened by zer0-x over 1 year ago

#857 - Fix tiny transcribe() docstring typo

Pull Request - State: closed - Opened by adamreis over 1 year ago

#845 - commented & deleted a few lines

Pull Request - State: closed - Opened by joe-bor over 1 year ago

#839 - Support batch-dimension in log_mel_spectogram

Pull Request - State: closed - Opened by HennerM over 1 year ago

#831 - *Fix catastrophic timestamp drifting from negative duration via clamping*

Pull Request - State: open - Opened by m-bain over 1 year ago - 3 comments

#812 - Use ndimage.median_filter instead of signal.medfilter

Pull Request - State: closed - Opened by mu4farooqi over 1 year ago - 3 comments

#811 - Use ndimage.median_filter instead of signal.medfilter

Pull Request - State: closed - Opened by mu4farooqi over 1 year ago

#804 - Add info about WER and which way it goes

Pull Request - State: closed - Opened by mikkovedru over 1 year ago

#789 - Add cli for downloading models

Pull Request - State: closed - Opened by nezhar over 1 year ago - 2 comments

#731 - Add `no_speech_threshold` and `logprob_threshold` to DecodingOptions

Pull Request - State: closed - Opened by MatthiasReumann over 1 year ago - 2 comments

#727 - Models are smaller?

Pull Request - State: closed - Opened by Abdurrafey-Siddiqui over 1 year ago - 1 comment

#681 - Add github action to automatically push to pypi on Release x.y.z commit

Pull Request - State: closed - Opened by rom1504 over 1 year ago - 7 comments

#676 - Improve language detection

Pull Request - State: open - Opened by PetrosVav over 1 year ago

#670 - verbose print catch UnicodeEncodeError

Pull Request - State: closed - Opened by simon300000 over 1 year ago - 2 comments

#659 - Fix bug where mm is mistakenly replaced with hmm in e.g. 20mm

Pull Request - State: closed - Opened by HennerM over 1 year ago - 1 comment

#630 - Closing file after reading

Pull Request - State: closed - Opened by paulharter over 1 year ago - 2 comments

#627 - Allows Whisper AI to be uploaded to pypi as a package

Pull Request - State: closed - Opened by zackees over 1 year ago - 1 comment

#561 - Fix `compression_ratio` function

Pull Request - State: closed - Opened by jumon over 1 year ago - 3 comments

#532 - Suppress non-timestamp tokens at the begging

Pull Request - State: closed - Opened by jumon over 1 year ago - 1 comment

#495 - Fix decoding error when mel is not given as encoded tensor

Pull Request - State: closed - Opened by TeemuSo over 1 year ago

#468 - [README] Add section on 🤗 Transformers

Pull Request - State: closed - Opened by sanchit-gandhi over 1 year ago - 4 comments

#401 - Update Hebrew language code to he per IANA registry

Pull Request - State: closed - Opened by altryne over 1 year ago - 2 comments

#399 - Disabled '...' from being generated, since it often gets generated

Pull Request - State: closed - Opened by shervinemami over 1 year ago - 3 comments

#382 - Uses MPS (Mac acceleration) by default when available

Pull Request - State: open - Opened by dwarkeshsp over 1 year ago - 52 comments

#370 - Fix attention caching to make transcription run 30% faster

Pull Request - State: closed - Opened by vickianand over 1 year ago - 6 comments

#333 - Added --output option

Pull Request - State: closed - Opened by Aaryan369 over 1 year ago - 5 comments

#299 - Bound end timestamps by length of audio input

Pull Request - State: open - Opened by isaacOnline over 1 year ago

#228 - Add CSV formatted output in transcript, using integer start/end times in milliseconds.

Pull Request - State: closed - Opened by NielsMayer over 1 year ago - 8 comments

#224 - timestamp should come after end of segment

Pull Request - State: closed - Opened by taylorchu over 1 year ago - 2 comments

#219 - Fix timestamps and strip extraneous whitespace in WebVTT output

Pull Request - State: closed - Opened by tomstuart over 1 year ago

#187 - Add Replicate demo and API 

Pull Request - State: closed - Opened by chenxwh almost 2 years ago - 1 comment

#141 - Use PyTorch as logits transpose for ONNX support

Pull Request - State: closed - Opened by mgoin almost 2 years ago - 9 comments