This project is modded from FunClip project, built with ASR (Automatic Speech Recognition), speaker identification, SRT editing, and LLM-based summarization capabilities. It integrates Gradio as the user interface, providing an interactive and easy-to-use platform.

简体中文 / English

本项目基于开源项目 FunClip 进行修改，集成了自动语音识别 (ASR)、说话人分离、SRT 字幕编辑以及基于 LLM 的总结功能。项目使用 Gradio 提供了一个直观易用的用户界面。

Update: Added support for GPU inference for both Docker/local deployment. Docker-GPU deployment Check This

This project builds upon the open-source FunClip by Alibaba DAMO Academy. I modded the functionality to include:

• ASR Summarization using LLMs (OpenAI GPT, custom API).
• Dynamic SRT Replacement with speaker mapping.
• Deployment Ready using Docker for production environments.

1. Automatic Speech Recognition (ASR):

   • Supports video and audio inputs.
   • Outputs text and SRT subtitles.

2. Speaker Identification (SD):

   • Identifies and differentiates speakers in multi-speaker audio/video.

3. SRT Subtitle Editing:

   • Replace speaker identifiers with user-defined names.

4. LLM Summarization:

   • Summarize ASR results using GPT-based models.
   • Allows custom API configurations.

5. Deployment Options:

   • Lightweight Docker container for production.
   • Python environment for development/testing.

• Docker (for containerized deployment)
• Python 3.9+ (for manual deployment)

See the requirements.txt file

Run the following command to build the Docker image:

docker build -t audio-processor:latest .

Use the following docker-compose.yml file for deployment:

version: '3.8' services: audio-processor: image: audio-processor:latest # The image you built container_name: audio-processor ports: - "7860:7860" volumes: - ./.env:/app/.env # Map the .env file working_dir: /app restart: unless-stopped

Run the deployment:

docker-compose up -d

The Gradio interface will be available at:
http://localhost:7860

1. Clone the repository:

   git clone https://github.com/MotorBottle/Audio-Processor.git cd audio-processor

2. Install dependencies:

   python3 -m venv .venv source .venv/bin/activate pip install --no-cache-dir -r requirements.txt

3. Ensure FFmpeg is installed(for Mac use brew):

   sudo apt-get update sudo apt-get install -y ffmpeg

Use the following command:

python funclip/launch.py --listen

The Gradio interface will be available at:
http://localhost:7860

Default user name: motor

Default passwd: admin

All credentials and API configurations can be stored in a .env file.

Example .env file:

USERNAME=motor PASSWORD=admin OPENAI_API_KEY=your_openai_key OPENAI_API_BASE=https://your-custom-api.com

1. Upload audio or video files.
2. Perform ASR Recognition or Speaker Differentiation.
3. Edit speaker names in the generated SRT subtitles.
4. Use the LLM Summarization feature to analyze and summarize the ASR text.

This project is released under the MIT License. Contributions are welcome!

For the original FunClip repository, visit:
FunClip on GitHub