Installation¶
Claw-R1 uses veRL as its training backend and Ray for distributed execution.
Prerequisites¶
| Requirement | Version | Notes |
|---|---|---|
| Python | 3.10+ | Conda recommended |
| CUDA | 11.8+ | Required for GPU training |
| Conda / Mamba | latest | For environment isolation |
| Git | any | For cloning repositories |
Step 1 — Create the Environment¶
Step 2 — Install veRL¶
Claw-R1 requires the nightly version of veRL installed from source:
Step 3 — Install Claw-R1¶
Step 4 — Install Ray and FastAPI¶
Verify Installation¶
Run the following checks to ensure everything is installed correctly:
# Check veRL
python -c "import verl; print('veRL:', verl.__version__)"
# Check Ray
python -c "import ray; print('Ray:', ray.__version__)"
# Check Claw-R1 gateway
python -m claw_r1.gateway.gateway --help
GPU Memory
The Gateway Server is CPU-only and lightweight. Rollout workers (vLLM) and Training workers require separate GPU pools. See Async Training for multi-GPU configuration.
What's Next¶
- Quick Start — run your first training loop
- Configuration Reference — customize the setup for your hardware