ChaoticByte/audio-summarize
Archived
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Clarify that only english summarization is supported at the moment, pin it in the code

Browse source
This commit is contained in:
ChaoticByte 2024-08-16 20:18:47 +02:00
parent f83043921a
commit f74b756db6
No known key found for this signature in database
2 changed files with 17 additions and 16 deletions
Download patch file Download diff file Expand all files Collapse all files

16
README.md
View file

@ -2,6 +2,10 @@
An audio summarizer that glues together [faster-whisper](https://github.com/SYSTRAN/faster-whisper) and [BART](https://huggingface.co/facebook/bart-large-cnn). An audio summarizer that glues together [faster-whisper](https://github.com/SYSTRAN/faster-whisper) and [BART](https://huggingface.co/facebook/bart-large-cnn).
## Supported Languages
Only English summarization is supported.
## Dependencies ## Dependencies
- Python 3 (tested: 3.12) - Python 3 (tested: 3.12)
@ -24,17 +28,15 @@ pip3 install -r requirements.txt
### Usage ### Usage
``` ```
./audio-summarize.py -i filepath -o filepath ./audio-summarize.py -i filepath -o filepath [-m name]
[--summin n] [--summax n] [--segmax n] [--summin n] [--summax n] [--segmax n]
[--lang lang] [-m name]
options: options:
-h, --help show this help message and exit -h, --help show this help message and exit
--summin n The minimum lenght of a segment summary [10, min: 5] --summin n The minimum lenght of a segment summary [10] (min: 5)
--summax n The maximum lenght of a segment summary [90, min: 5] --summax n The maximum lenght of a segment summary [90] (min: 5)
--segmax n The maximum number of tokens per segment [375, 5 - 500] --segmax n The maximum number of tokens per segment [375] (5 - 500)
--lang lang The language of the audio source ['en'] -m name The name of the whisper model to be used [small.en]
-m name The name of the whisper model to be used ['small.en']
-i filepath The path to the media file -i filepath The path to the media file
-o filepath Where to save the output text to -o filepath Where to save the output text to
``` ```

15
audio-summarize.py
View file

@ -20,12 +20,12 @@ from transformers import pipeline
# Transcription # Transcription
def transcribe(model_name: str, audio_file: str, language: str) -> str: def transcribe(model_name: str, audio_file: str) -> str:
'''Transcribe the media using faster-whisper''' '''Transcribe the media using faster-whisper'''
t_chunks = [] t_chunks = []
print("* Loading model ", end="", flush=True) print("* Loading model ", end="", flush=True)
model = WhisperModel(model_name, device="auto", compute_type="int8") model = WhisperModel(model_name, device="auto", compute_type="int8")
segments, _ = model.transcribe(audio_file, language=language, beam_size=5, condition_on_previous_text=False) segments, _ = model.transcribe(audio_file, language="en", beam_size=5, condition_on_previous_text=False)
print() print()
print("* Transcribing audio ", end="", flush=True) print("* Transcribing audio ", end="", flush=True)
for s in segments: for s in segments:
@ -67,11 +67,10 @@ def summarize(chunks: List[str], summary_min: int, summary_max: int) -> str:
if __name__ == "__main__": if __name__ == "__main__":
# parse commandline arguments # parse commandline arguments
argp = ArgumentParser() argp = ArgumentParser()
argp.add_argument("--summin", metavar="n", type=int, default=10, help="The minimum lenght of a segment summary [10, min: 5]") argp.add_argument("--summin", metavar="n", type=int, default=10, help="The minimum lenght of a segment summary [10] (min: 5)")
argp.add_argument("--summax", metavar="n", type=int, default=90, help="The maximum lenght of a segment summary [90, min: 5]") argp.add_argument("--summax", metavar="n", type=int, default=90, help="The maximum lenght of a segment summary [90] (min: 5)")
argp.add_argument("--segmax", metavar="n", type=int, default=375, help="The maximum number of tokens per segment [375, 5 - 500]") argp.add_argument("--segmax", metavar="n", type=int, default=375, help="The maximum number of tokens per segment [375] (5 - 500)")
argp.add_argument("--lang", metavar="lang", type=str, default="en", help="The language of the audio source ['en']") argp.add_argument("-m", metavar="name", type=str, default="small.en", help="The name of the whisper model to be used [small.en]")
argp.add_argument("-m", metavar="name", type=str, default="small.en", help="The name of the whisper model to be used ['small.en']")
argp.add_argument("-i", required=True, metavar="filepath", type=Path, help="The path to the media file") argp.add_argument("-i", required=True, metavar="filepath", type=Path, help="The path to the media file")
argp.add_argument("-o", required=True, metavar="filepath", type=Path, help="Where to save the output text to") argp.add_argument("-o", required=True, metavar="filepath", type=Path, help="Where to save the output text to")
args = argp.parse_args() args = argp.parse_args()
@ -80,7 +79,7 @@ if __name__ == "__main__":
args.summax = max(5, args.summax) args.summax = max(5, args.summax)
args.segmax = max(5, min(args.segmax, 500)) args.segmax = max(5, min(args.segmax, 500))
# transcribe # transcribe
text = transcribe(args.m, args.i, args.lang).strip() text = transcribe(args.m, args.i).strip()
# split up into semantic segments & summarize # split up into semantic segments & summarize
chunks = split_text(text, args.segmax) chunks = split_text(text, args.segmax)
summary = summarize(chunks, args.summin, args.summax) summary = summarize(chunks, args.summin, args.summax)