Speech-to-Text using ggerganov/whisper.cpp for GitHub Action. High-performance inference of OpenAI Whisper automatic speech recognition (ASR) model.
See action.yml for more detailed information.
| Variable | Description | Default |
|---|---|---|
| model | public whisper model. (available: small, medium and large) | small |
| audio_path | Audio Path. | |
| output_folder | output folder. | |
| output_format | output format, support txt, srt, csv. | txt |
| output_filename | output filename. | |
| debug | enable debug mode. | |
| print_progress | print progress. | true |
| print_segment | print segment. | |
| youtube_url | youtube url | |
| translate | translate from source language to english. | false |
| cut_silences | cut silences. | false |
| prompt | initial prompt text. |
Donwload Youtube video and transcript it.
jobs:
youtube-eng-video:
name: transcript english video
runs-on: ubuntu-latest
steps:
- name: checkout
uses: actions/checkout@v3
- name: speech to text
uses: appleboy/whisper-action@v0.1.1
with:
model: small
youtube_url: https://www.youtube.com/watch?v=pTCxXZh6VyE
output_format: srt
output_folder: youtube
print_segment: true
debug: true
- name: git push changes
uses: appleboy/git-push-action@v0.0.2
with:
branch: main
commit: true
commit_message: "[skip ci] Upload changes"
remote: git@github.com:appleboy/whisper-action.git
ssh_key: ${{ secrets.DEPLOY_KEY }}
rebase: trueSee the output file in youtube folder.