A Docker tool for transcribing video files with spoken words. The intended use of transcriptions is as input to other text analysis tools, such as LLM summarization or indexing.
This is mostly a Docker wrapper around OpenAI's Whisper.
- Can handle multiple video files, concatenating videos in order
- Extracts audio from video files
- Transcribes audio using OpenAI's Whisper, output is Whisper transcript format.
- Outputs transcription in JSON and plain text formats
- Generates a manifest file with metadata and checksums
- GPU acceleration
- Docker
To build the Docker image, run:
docker build -t transcription-tools-gpu ./transcription
docker run --gpus all -ti -v "$PWD/transcription/data:/usr/data" transcription-tools-gpu --input /usr/data/video.mkv --output-dir /usr/data --model large-v3
Will output:
data/video1.manifest.json
data/video1.transcript.json
data/video1.transcript.txt