cbf2b022a5
Replace YAMNet model for panns-inference
...
YAMNet model was causing issues, as loading a pytorch framework model, when the tensorflow based YAMNet is loaded in the same environment already, caused segmentation fault
2025-02-26 17:36:27 +05:30
f2011b4408
endpoint : /api/separate
...
- Define new endpoint /api/separate, to separate music using demucs, params: file_uuid
- Replace original file with vocals.wav while retaining original filename
- Move all other files to file_path/sources/
2025-02-26 14:23:08 +05:30
b0939cb5b8
fix bug: return original sample rate
2025-02-26 13:56:59 +05:30
7672817ff3
return sr with response of /api/upload/
2025-02-26 13:37:08 +05:30
a80a851c80
endpoint: /api/resample
...
- Define new endpoint /api/resample, which takes param : file_uuid, and sr.
- Add Exception Handling for celery tasks
- Correct error message for trim_audio_task
2025-02-26 03:05:41 +05:30
8356394b8a
remove redundant code, improve get_audio_file_path()
...
- Check for file_uuid in request within function
- Check for existence of file within function
- Return appropriate status codes
2025-02-26 02:39:27 +05:30
f71d30732f
endpoint: /api/trim
...
- Define new api endpoint /api/trim, params: file_uuid, overwrites exisiting file on server with trimmed audio.
- Remove print messages in freqsplit/postprocessing/audio_writer.export_audio()
- Add new file api/utils.py for commonly used functions
2025-02-26 01:59:33 +05:30
1799713d3b
endpoint: /api/normalize
...
-Add new endpoint /api/normalize, params: file_uuid, overwrites exisisting file on server with normalized audio.
-Remove print statements from src/freqsplit/postprocessing/audio_writer
2025-02-26 00:19:23 +05:30
2256989fda
check file format in /api/upload/
2025-02-25 23:34:54 +05:30
9bc3a90cf9
remove unused imports
2025-02-25 23:00:27 +05:30
4f86f7a94b
/api/upload: synchronously uploads audio, returns file_uuid, classifies audio, returns audio_class
2025-02-25 22:53:49 +05:30
26367e3497
bugfix: use parameter mono only if passed explicitly
...
Caused the tests/test_separation to fail.
2025-02-25 20:53:39 +05:30
3074084ac1
refactor code, use package libraries for freqsplit/preprocessing/classify
...
- add additional params: sr=None and mono=False for freqsplit/input/file_reader/read_audio()
- remove loading audio using librosa in freqsplit/preprocessing/classify/classify_audio()
- add error handling, if the sr is not 16Khz
- raise RuntimeError, if YAMNet model fails
- update tests/test_preprocessing
2025-02-25 20:13:02 +05:30
825da48712
use celery for asynchronous tasks, rename session_uuid -> file_uuid
2025-02-25 19:30:58 +05:30
0d35e4e071
use session id's for uploads, return it as a response
...
also removed classify for api/upload temporarily
2025-02-25 19:07:15 +05:30
b51edcf173
rename package: freq-split-enhance -> freqsplit
2025-02-25 19:05:37 +05:30
9561f973aa
Merge pull request #28 from joelmathewthomas/update/readme
...
Update/readme
2025-02-25 03:46:08 +05:30
404150c834
Update README.md
2025-02-25 03:45:13 +05:30
1943cc006c
Update README.md
2025-02-25 03:44:18 +05:30
4246e3f914
Merge pull request #27 from joelmathewthomas/feature/api
...
Feature/api
2025-02-25 03:36:49 +05:30
e61ae413af
return audio_class with response for api/upload endpoint
2025-02-25 03:23:59 +05:30
fc03a50ab2
update deps
2025-02-25 03:19:21 +05:30
a8edb05d28
update pyproject.toml
2025-02-25 03:15:47 +05:30
0e0e765b6e
install freqsplit as a package
2025-02-25 03:10:44 +05:30
076e3779ef
define new route api/upload to upload audio
2025-02-25 02:24:05 +05:30
c2680bdb9d
Merge pull request #26 from joelmathewthomas/update/deps
...
update deps
2025-02-25 01:46:47 +05:30
e641702972
update deps
2025-02-25 01:45:33 +05:30
6afb27f294
Merge pull request #25 from joelmathewthomas/update/readme
...
Update README.md
2025-02-25 01:44:37 +05:30
030eb4710e
Update README.md
2025-02-25 01:42:39 +05:30
436f2d5b22
Merge pull request #24 from joelmathewthomas/feature/deepfilternet
...
Feature/deepfilternet
2025-02-24 21:08:43 +05:30
10921cdad3
disabled cuda for yamnet model, and removed multi-env setup scripts, revert back to single env
2025-02-24 20:49:16 +05:30
90d3076e65
create test for noisereduce
2025-02-24 20:17:00 +05:30
55d10b73e9
create deepfilternet_wrapper
2025-02-24 20:16:36 +05:30
05f9c877e3
update deps
2025-02-24 20:16:12 +05:30
e7b142504c
Merge pull request #23 from joelmathewthomas/feature/asteroid-wrapper
...
Feature/asteroid wrapper
2025-01-29 00:51:33 +05:30
f2e41dc7ef
update deps
2025-01-29 00:47:23 +05:30
489a37c007
combine separation tests
...
combined separation tests, because classify_audio would fail when other models ran on GPU
2025-01-29 00:28:54 +05:30
7f973db1d0
add test case to test mpariente/ConvTasNet_WHAMsepcleanwrapper
2025-01-28 23:51:21 +05:30
fae60a8120
Fix: Corrected shape handling for separated sources, extracting each source as a separate 1D array
2025-01-28 23:34:08 +05:30
9e7c01c19e
speech separation works, but ends up on either channels of same audio
2025-01-28 23:19:52 +05:30
88ebfb66f6
change model
2025-01-28 22:39:45 +05:30
2982aaaafd
add postprocessing package and create module for audio export
2025-01-28 20:33:57 +05:30
3914b2c5ba
remove unused parameter
2025-01-28 20:33:28 +05:30
7a7c27caa3
create ConvTasNet wrapper using asteroid
2025-01-28 20:20:48 +05:30
8e3e7dd062
Merge pull request #21 from joelmathewthomas/bugfix/scripts
...
Bugfix/scripts
2025-01-26 18:28:28 +05:30
a041e7f0ef
add script which can be sourced to setup the current SHELL environment
2025-01-26 18:11:19 +05:30
e63513d336
remove site-packages.pth
2025-01-26 18:10:53 +05:30
8900fabe57
correctly set path in site-packages.pth
2025-01-26 17:10:03 +05:30
0fa92c2656
Merge pull request #20 from joelmathewthomas/refactor/venv
...
refactor/venv
2025-01-26 16:46:52 +05:30
5975097ed5
update deps
2025-01-26 16:41:43 +05:30
Joel Mathew Thomas