Getting Started
Installation
Install ThinkRL and start training RLHF models in minutes.
Requirements
- Python 3.9+
- PyTorch 2.0+
- CUDA 11.8+ (for GPU training)
- 8GB+ GPU VRAM (16GB+ recommended)
Quick Installation
The recommended way to install ThinkRL right now is from source:
# Clone the repository
git clone https://github.com/ellanorai/ThinkRL.git
cd ThinkRL
# Create and activate virtual environment
python -m venv .venv
source .venv/bin/activate # On Unix
# Or: .venv\Scripts\activate on Windows
# Install in editable mode
pip install -e .
# Or with all extras (vLLM, DeepSpeed, etc.)
pip install -e ".[all]"Development Installation
For contributors or to access the latest features:
# Clone the repository
git clone https://github.com/ellanorai/ThinkRL.git
# Navigate to the directory
cd ThinkRL
# Install in development mode
pip install -e .
# Install development dependencies
pip install -e .[dev]Optional Dependencies
vLLM Support
pip install -e ".[vllm]"Enables 10x faster generation during RLHF training.
DeepSpeed Support
pip install -e ".[deepspeed]"Enables ZeRO optimization for large models.
Verify Installation
Verify that ThinkRL is installed correctly:
import thinkrl
print(thinkrl.__version__)
# Expected output: 1.0.0