README.md

metadata

title: WhisperPyanoteLLM
emoji: 📉
colorFrom: indigo
colorTo: green
sdk: docker
pinned: false
license: apache-2.0

Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

WhisperPyanoteLLM

A FastAPI-based app for speaker diarization and transcription using Whisper and PyAnnote, with LLM-powered summarization.

Features

• Speaker diarization with pyannote.audio
• Transcription with OpenAI Whisper
• Summarization with Together LLM
• REST API for video/audio upload and processing

Quick Start (Development)

1. Clone the repository:

   git clone <your-repo-url>
   cd WhisperPyanoteLLM
   

2. Create a .env file:

   HF_TOKEN=your_huggingface_token
   TOGETHER_API_KEY=your_together_api_key
   NGROK_AUTH_TOKEN=your_ngrok_token
   

3. Install dependencies:

   pip install -r requirements.txt
   

4. Run the app:

   uvicorn app:app --reload --port 8300
   

5. Access the API:

   • Health check: http://localhost:8300/health
   • Upload endpoint: /upload_video/

---

Production (Docker)

1. Create a .env.prod file:

   HF_TOKEN=your_huggingface_token
   TOGETHER_API_KEY=your_together_api_key
   NGROK_AUTH_TOKEN=your_ngrok_token
   

2. Build the Docker image:

   docker build -t whisperpyanote .
   

3. Run the Docker container:

   docker run --env-file .env.prod -p 8300:8300 whisperpyanote
   

4. Access the API:

   • Health check: http://localhost:8300/health
   • Upload endpoint: /upload_video/

---

Notes

• Make sure your .env and .env.prod files are not committed to version control.
• For best performance, run on a machine with a CUDA-enabled GPU.
• For more details, see the code and comments in app.py.

---

License

Apache-2.0