/esm3-exostack-cpu

This repository integrates ESM, EvolutionaryScale's protein language models, with the ExoStack platform to enable scalable AI-driven protein analysis, inference, and deployment.

🧬 ESM + ExoStack Integration

This repository integrates ESM, EvolutionaryScale's protein language models, with the ExoStack platform to enable scalable AI-driven protein analysis, inference, and deployment.

📁 Project Structure


ESM/
├── esm/              # ESM protein language model (submodule)
│   └── esm/          # Core ESM3 model components
├── exostack/         # ExoStack framework for orchestration
│   ├── exo\_agent/    # Custom agent with esm3\_executor
├── models/           # Pretrained / fine-tuned model artifacts
├── logs/             # Logs from training and inference
└── esm3env/          # Python virtual environment (not committed)

⚙️ Setup Instructions

1. Clone the Repository

git clone --recurse-submodules https://github.com/yourusername/ESM
cd ESM

⚠️ Note: The esm/ folder is a Git submodule. Use --recurse-submodules to fetch it correctly.

2. Create and Activate a Virtual Environment

python3 -m venv esm3env
source esm3env/bin/activate

3. Install Required Dependencies

# ESM requirements
cd esm
pip install -r requirements.txt
cd ..

# ExoStack requirements
cd exostack
pip install -r requirements.txt
cd ..

4. Run ESM3 Integration via ExoStack

cd exostack
python exo_agent/esm3_executor.py

🚀 Features

  • ⚡ ESM3-powered protein language modeling
  • 🔁 Integrated inference pipeline with ExoStack agents
  • 📦 Docker-ready setup
  • 🌐 Scalable K8s-compatible microservices
  • 📈 Model serving + logging

🧪 Testing

cd exostack
pytest

🧠 Example Use Case

You can modify exo_agent/esm3_executor.py to run predictions using ESM3 and route outputs via ExoStack’s pub-sub or task queue model.

📝 Contributing

Contributions welcome! Please open issues or PRs.

📄 License