ericwang2006/funclip repository overview

This Docker image is specifically created for the FunClip project⁠ hosted on GitHub.

⁠What is FunClip?

FunClip⁠ is a fully open-source, locally deployable automated video clipping tool. It utilizes the open-source FunASR Paraformer series models developed by Alibaba TONGYI speech lab to perform speech recognition on videos. Users can then freely select text segments or specific speakers from the transcription results and click the "Clip" button to extract the corresponding video clip.

⁠Usage

⁠Run with GPU:

docker run -d -p 7860:7860 \
--name funclip \
--gpus all \
--restart always \
ericwang2006/funclip

⁠Run without GPU:

docker run -d -p 7860:7860 \
--name funclip \
--restart always \
ericwang2006/funclip

⁠Run in Command-Line Mode

You can also run FunClip directly in command-line mode for single video processing without launching the web UI. This method is useful for batch scripts or automation:

docker run --rm \
-v ./modelscope:/root/.cache/modelscope \
-v ./output:/output \
ericwang2006/funclip \
python funclip/videoclipper.py --stage 1 --file /output/example.mp4 --output_dir /output

In this command:

• example.mp4 is the input video file, located in the ./output directory on the host.
• ./output is the directory where the transcription results and clips will be saved.
• ./modelscope is used to cache downloaded models locally to avoid re-downloading them each time.

Make sure example.mp4 exists in your ./output folder before running the command.

⁠Wait for Initialization

After the container starts, please wait a few minutes. The application will automatically download the necessary speech recognition models on the first run. Once the models are downloaded and the service is fully initialized, you can access the web interface via:

http://<your-host-ip>:7860

⁠Usage Instructions

For detailed instructions on how to use FunClip, please refer to the Use FunClip⁠ section in the official repository.

⁠Supported Architectures

This Docker image supports the following system architectures:

• amd64
• arm64