983519e629
Add compatibility with python 3.5 issue( #1021 ) ( #1041 )
...
Added format for python3.5
2022-06-29 18:00:40 +03:00
859420809b
Add extra comments
2022-05-26 11:45:16 +02:00
4209f3a9fe
Vosk model loader( #871 ) ( #941 )
...
* methods get_model_by_name, get_model_by_lang, get_model were added into the model class
* importing modules changed to using components; introduced constant MODELS_HOME_DIR; simplified code
* added new model folders into init; changed samples and transcriber bin for new mode loader
* changed back in cli.py lang arg to args.lang
* added 3 directories instead of 1 to check for models
* cli.py: added 3 args instead of 1 for model; __init__.py: changed script get_model_path for run get_model_by_name/lang inside current directory
* deleted default env var
* cli.py: changed arg_name; __init__.py: changed const name, changed model loading only for last directory
* deleted unused method
* changed by_name, by_lang methods, added download_model method
* deleted env variable initialization
* deleted print()
* deteled unused modules
* added progress_bar, added folder AppData/Local/vosk for model search
* changed download_model methond; added my_hook method
2022-05-24 21:06:13 +03:00
def8c93711
Remove semicolon
2022-05-24 15:23:03 +03:00
7b7d814484
Introduce incremental decoder with confidences in partial results
2022-04-07 01:07:47 +02:00
a57a84f90e
Refactor GPU API to hide the ID and keep it closer to CPU recognizer
2022-03-03 21:09:09 +01:00
79b8395be0
Add NLSML output
2022-02-03 23:08:09 +01:00
d2c11a611f
Read list of files from arguments
2022-01-30 22:57:36 +01:00
72bf210164
Put the demo into main folder
2021-12-24 01:07:38 +01:00
cb0f8e6411
Per-stream wait API
2021-12-23 22:34:47 +01:00
848b2dc753
Expose results in Python
2021-12-17 22:57:00 +01:00
60f0396fe0
Reset lattice on endpoint
2021-12-17 01:13:09 +01:00
344e137a61
Decoding works, results are empty yet
2021-12-13 01:21:59 +01:00
6977be7fb7
Batch recognizer draft
2021-12-12 21:37:44 +01:00
2349e66a97
Subtitles require word times ( #607 )
...
This is a port of the recent addition of commit 7ccf743
2021-06-27 18:06:11 +03:00
02ef49f67e
Allow Saving WebVTT ( #605 )
...
This patch is a small extension to the WebVTT example which allows to
directly save the WebVTT output to a file like this:
./test_webvtt.py test.wav out.vtt
2021-06-24 19:35:30 +03:00
7cdf8f1d03
Add Python WebVTT Example ( #601 )
...
This patch adds an example for using webvtt-py to generate WebVTT files
from Vosk output. This is similar to the SRT example but still very
useful for generating an example video subtitle usable in web contexts.
2021-06-23 01:16:33 +03:00
7ccf743bb6
SRT requires word times
2021-06-10 10:37:15 +02:00
75bedfe06d
Add a method to show/hide words and their times
2021-06-07 01:04:37 +02:00
6aa5af7640
Add reset test
2021-05-26 21:22:07 +02:00
499b2f183a
Introduce new API to set speaker model to already initialized recognizer.
...
Introduce a method to reset recognizer results to start from scratch without
computation of the result.
2021-05-26 00:46:32 +02:00
f8189685e5
Add max alternatives output
2021-05-19 18:47:25 +02:00
a5a3697b7c
Copy data before queue, original data can be destroyed.
...
Fixes issue #444
Thanks to Alexander Zatvornitsky
2021-03-01 21:25:41 +01:00
c6119c4835
test_microphone: AcceptWavform wants bytes
2021-02-08 00:38:28 +02:00
6f2d6d0d69
Proper microphone recognizer with the queue
2021-01-09 23:16:35 +01:00
08c35e84f3
Update demo with spk vector check
2020-12-23 22:21:12 +01:00
dc3d03d742
Make sure we have result field in json
2020-11-29 19:19:50 +01:00
746ff47757
Split long lines in subtitles
2020-11-03 15:13:16 +01:00
7af3e9a334
Add srt example
2020-10-07 13:43:09 +02:00
57cc474c9f
Build bigram language model from grammars
2020-10-04 23:42:50 +02:00
f97383c17f
Update models location
2020-09-22 00:00:24 +02:00
41035485db
Fix x-vectors, now they actually work. Requires new version spk-model-0.4 with whitening transform matrix
2020-09-21 23:31:30 +02:00
1c7b94757d
Don't attempt to resize to empty matrix
2020-07-31 16:45:05 +02:00
722b09eaa4
Ignore words missing in the vocabulary
2020-07-08 01:35:50 +02:00
b81f69d407
Avoid overflow. See issue #128
2020-06-06 10:43:44 +02:00
b1e775c67b
Better speaker identification without silence frames
2020-05-31 23:46:42 +02:00
37fbe1a52b
OSX build
2020-05-08 09:55:15 +03:00
af11bb2361
No need for external dependencies
2020-05-07 21:48:03 +02:00
8da8697c1e
Add ffmpeg test
2020-05-07 21:07:17 +02:00
80219066e9
Expose verbose level in the API
2020-05-01 19:02:57 +02:00
d75bb36131
Fix model download path
2020-04-30 11:04:35 +02:00
86caf526f5
New model layout and model config file with beams for decoding
2020-04-28 01:06:41 +02:00
14b2c13ed6
Fixes folder name instruction for python model on microphone test
2020-03-25 01:01:23 -03:00
095fac1de4
Small unittest
2020-02-25 21:45:27 +01:00
8913b93703
Fix message
2020-02-23 16:05:34 +00:00
f66f7a9290
Added microphone test
2020-02-21 00:08:50 +01:00
378fa3499a
Downsample by default and also check wav file encoding in examples. Thanks to
...
dtreskunov. See for details
https://github.com/alphacep/vosk-api/pull/23
2020-02-18 11:45:50 +01:00
01eeec498d
Fix memory access issues
2020-02-17 09:42:50 +00:00
d422dfbddb
Message update
2020-02-16 17:21:38 +01:00
aa91ccf68b
Added speaker identification
2020-02-16 17:04:37 +01:00
vadimdddd