31 lines
903 B
Markdown
31 lines
903 B
Markdown
# 🗣️ ✍️ 💾
|
|
|
|
STT terminal utility based on openai-whisper and pyaudio
|
|
|
|
```
|
|
git clone --recurse-submodules <git-repo>
|
|
```
|
|
|
|
```
|
|
usage: ./transcriptum.sh [action]
|
|
where action can be: [install, clean, run]
|
|
```
|
|
|
|
```
|
|
usage: transcribe.py [-h] [--model {tiny,base,small,medium,large}] [--rms RMS]
|
|
[--record_timeout RECORD_TIMEOUT]
|
|
[--phrase_timeout PHRASE_TIMEOUT] [--dynamic_threshold]
|
|
|
|
TRANSCRIPTUM
|
|
|
|
options:
|
|
-h, --help show this help message and exit
|
|
--model {tiny,base,small,medium,large}
|
|
Whisper model
|
|
--rms RMS RMS (energy) threshold for microphone to detect
|
|
--record_timeout RECORD_TIMEOUT
|
|
Timeout for the microphone recording
|
|
--phrase_timeout PHRASE_TIMEOUT
|
|
Silence timeout between phrases
|
|
--dynamic_threshold Use dynamic rms threshold?
|
|
``` |