heimoshuiyu/faster-whisper
1
0
Fork 0
Code Issues Pull Requests 1 Packages Projects Releases Wiki Activity
88 Commits 2 Branches 0 Tags
a5d03e55fa4f977ed58757055a05f9ab0d5e749f
Commit Graph

60 Commits

Author SHA1 Message Date
Guillaume Klein
a5d03e55fa Prevent out of range error in method split_tokens_on_unicode (#111) 2023-04-04 10:51:14 +02:00
Guillaume Klein
9fa1989073 Revert "Prevent out of range error in method split_tokens_on_unicode" 
This reverts commit 36160c1e7e.
 2023-04-04 10:25:41 +02:00
Guillaume Klein
36160c1e7e Prevent out of range error in method split_tokens_on_unicode 2023-04-04 10:17:56 +02:00
Guillaume Klein
2f266eb844 Fix VAD index error when a predicted timestamps is too large (#107) 2023-04-03 19:34:54 +02:00
Guillaume Klein
19698c95f8 Support VAD filter (#95) 
* Support VAD filter

* Generalize function collect_samples

* Define AudioSegment class

* Only pass prompt and prefix to the first chunk

* Add dict argument vad_parameters

* Fix isort format

* Rename method

* Update README

* Add shortcut when the chunk offset is 0

* Reword readme

* Fix end property

* Concatenate the speech chunks

* Cleanup diff

* Increase default speech pad

* Update README

* Increase default speech pad
 2023-04-03 17:22:48 +02:00
palladium123
b4c1c57781 Added retrieval mechanism (avg_log_prob/no_speech_prob) (#103) 
* Added retrieval mechanism 

Added retrieval mechanism to retrieve avg_log_prob and no_speech_prob from the Transcribe method.

* Update transcribe.py

* Update transcribe.py

* Initial commit
 2023-04-03 16:56:35 +02:00
Guillaume Klein
f20bb258de Support separating the left and right audio channels (#97) 2023-04-03 11:22:43 +02:00
Guillaume Klein
1a968a4323 Pass prefix only to the first window 2023-04-01 09:27:20 +02:00
Guillaume Klein
d03383f902 Simplify reuse of the encoder output 2023-03-30 15:58:27 +02:00
Guillaume Klein
39fddba886 Suppress some special tokens when the default set is not used 2023-03-30 12:42:29 +02:00
Guillaume Klein
eda840f8ff Always disable the progress bar specific to snapshot_download 2023-03-29 12:11:24 +02:00
Guillaume Klein
0224400584 Add large-v1 model 2023-03-28 14:36:10 +02:00
Guillaume Klein
8246479fda Ignore the invalid audio frames (#82) 2023-03-27 10:19:22 +02:00
Guillaume Klein
e2705d11c9 Raise an explicit error message if the model size is invalid 2023-03-26 16:30:00 +02:00
Jordi Mas
f8d2fb169f Fix variable name reference (#77) 2023-03-25 10:00:59 +01:00
Guillaume Klein
a10732c74a Only download the required model files 2023-03-24 17:59:11 +01:00
Guillaume Klein
de7682a2f0 Automatically download converted models from the Hugging Face Hub (#70) 
* Automatically download converted models from the Hugging Face Hub

* Remove unused import

* Remove non needed requirements in dev mode

* Remove extra index URL when pip install in CI

* Allow downloading to a specific directory

* Update docstring

* Add argument to disable the progess bars

* Fix typo in docstring
 2023-03-24 10:55:55 +01:00
Guillaume Klein
523ae2180f Run the encoder only once for each 30-second window (#73) 2023-03-24 10:53:49 +01:00
Guillaume Klein
66efd02bd0 Run some automatic tests with GitHub Actions (#68) 2023-03-22 20:50:03 +01:00
Guillaume Klein
52264f2277 Fix typing for device_index argument 2023-03-22 13:51:12 +01:00
Guillaume Klein
c27c010f96 Ignore Unicode errors in input file metadata 2023-03-21 17:13:37 +01:00
Guillaume Klein
0ab8db2b37 Remove debug prints 2023-03-18 09:48:02 +01:00
Guillaume Klein
a70aac18ae Remove unused import 2023-03-18 09:47:02 +01:00
Guillaume Klein
d82be59d5f Fix unset attribute when using English-only models 2023-03-17 18:33:16 +01:00
Guillaume Klein
cce6b53e45 Fix incorrect attribute access 2023-03-16 10:32:36 +01:00
Guillaume Klein
2007adf0b5 Fix typing of words attribute 2023-03-15 17:49:07 +01:00
Guillaume Klein
ae9898f0d8 Include duration in AudioInfo structure 2023-03-15 15:30:29 +01:00
Guillaume Klein
c5f6b91b7d Port utility function format_timestamp 2023-03-15 15:27:20 +01:00
Guillaume Klein
eafb2c79a3 Add more typing annotations 2023-03-15 15:22:53 +01:00
Guillaume Klein
8bd013ea99 Add word-level timestamps (#43) 
* Add word-level timestamps

* Fix alignment between the segments and the lists of words

* Fix truncated words list when the replacement character is decoded

* Check for empty text_tokens

* Add usage example in the readme

* Update ctranslate2 to 3.9

* Skip empty segment

* Set typing for the new methods
 2023-03-15 15:02:28 +01:00
Guillaume Klein
3301dd9273 Make get_input a free function 2023-03-09 12:54:41 +01:00
Guillaume Klein
c52adaca90 Create a helper class Tokenizer 2023-03-09 12:53:49 +01:00
Guillaume Klein
f0a21ea916 Use a dict to represent intermediate segments 2023-03-09 11:53:55 +01:00
Guillaume Klein
6a84df400f Fix all_tokens handling 
See 38f2f4d99d
 2023-03-09 10:02:25 +01:00
Guillaume Klein
4176da0d68 Rename offset to seek to match the OpenAI implementation 2023-03-09 09:58:58 +01:00
Guillaume Klein
6b16b8a69c Pad the audio instead of the spectrogram 
See 919a713499
 2023-03-08 10:50:46 +01:00
Guillaume Klein
2646906596 Fix error in decode_audio for long audio inputs 2023-03-07 10:15:36 +01:00
Guillaume Klein
01ef12a6a0 Do not ignore last segment ending with one timestamp 
See eab8d920ed
 2023-03-07 10:05:04 +01:00
Guillaume Klein
469244a57d Update CTranslate2 to 3.8.0 2023-03-06 16:21:48 +01:00
Guillaume Klein
4a18adc382 Load the tokenizer from the model directory if it exists 2023-03-01 15:47:16 +01:00
Guillaume Klein
873992623c Accept the audio waveform as an input to transcribe() (#21) 2023-02-28 19:01:31 +01:00
Guillaume Klein
a4f1cc8f11 Add prefix parameter 2023-02-27 12:09:40 +01:00
Guillaume Klein
528aa3e784 Make threshold parameters optional 2023-02-27 11:32:03 +01:00
Guillaume Klein
f0add58bdc Add typing to constructor and transcribe method 2023-02-27 11:22:02 +01:00
Guillaume Klein
ef71be09ed Update CTranslate2 to 3.7.0 2023-02-23 11:18:58 +01:00
Guillaume Klein
d91365e321 Minor code simplification 2023-02-22 11:02:11 +01:00
Guillaume Klein
4b8237da1b Strip the leading space before computing the compression ratio 2023-02-22 10:28:04 +01:00
Guillaume Klein
e47e00910a Add length_penalty parameter and correctly compute the avg log prob 2023-02-22 10:27:38 +01:00
Guillaume Klein
f5c9f15c2c Check that the language code is valid 2023-02-21 12:10:54 +01:00
Guillaume Klein
e2094b6474 Reduce the maximum length when the prompt is longer than 448/2 2023-02-17 14:37:24 +01:00