Feature/faster whisper #2
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Contributor
ruokolt
commented
Jan 11, 2024
•
ruokolt
commented
- Replace OpenAI Whisper with Faster Whisper.
- Upgrade pyannote diarization pipeline from version 3.0 to 3.1
- Update readme and docs
- Add .lua version 20240111
Create the faster-whisper branch from personal fork
hsnfirooz
approved these changes
Jan 11, 2024
Comment on lines
+266
to
+294
| def convert_to_wav(input_file, tmp_dir): | ||
| """Pyannote diarization pipeline does handle resampling to ensure 16 kHz and | ||
| stereo/mono mixing. However, number of supported audio/video formats appears to be | ||
| limited and not listed in README. To be sure, we convert all files to .wav beforehand. | ||
|
|
||
| https://huggingface.co/pyannote/speaker-diarization-3.1 | ||
| """ | ||
|
|
||
| if str(input_file).lower().endswith(".wav"): | ||
| logger.info(f".. .. File is already in wav format: {input_file}") | ||
| return input_file | ||
|
|
||
| if not Path(input_file).is_file(): | ||
| logger.info(f".. .. File does not exist: {input_file}") | ||
| return None | ||
|
|
||
| converted_file = Path(tmp_dir) / Path(Path(input_file).name).with_suffix(".wav") | ||
| if Path(converted_file).is_file(): | ||
| logger.info(f".. .. Converted file {converted_file} already exists.") | ||
| return converted_file | ||
| try: | ||
| AudioSegment.from_file(input_file).export(converted_file, format="wav") | ||
| logger.info(f".. .. File converted to wav: {converted_file}") | ||
| return converted_file | ||
| except Exception as err: | ||
| logger.info(f".. .. Error while converting file: {err}") | ||
| return None | ||
|
|
||
|
|
There was a problem hiding this comment.
Not sure if we are using the same code for Kubernetes in the future or not, but just checking the suffix of a file is not enough.
How WhisperX and OpenAI Whisper did the format change were calling a subprocess with ffmpeg in the background. Look at here.
I suggest we can do the same, but it's not necessary at this point.
There was a problem hiding this comment.
Yes, the suffix check is hacky and should be replaced with a proper check in the future. ffmpeg in a subprocess works also for the conversion.
ruokolt
added a commit
that referenced
this pull request
Jan 12, 2024
- Replace OpenAI Whisper with Faster Whisper - Upgrade pyannote diarization pipeline from version 3.0 to 3.1 - Update readme and docs - Add .lua version 20240111
Closed
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.