TensorTrade is a powerful open-source Python framework for building, training, evaluating, and deploying robust trading algorithms using reinforcement learning. Currently in Beta (v1.0.4-dev1), it combines cutting-edge machine learning with algorithmic trading to create sophisticated trading strategies.
- Highly Composable: Build complex strategies from simple, reusable components
- Production Ready: Scale from single CPU to distributed HPC environments
- ML Integration: Seamless integration with popular ML libraries (numpy, pandas, gym, keras, tensorflow)
- Fast Experimentation: Rapid prototyping and testing of trading strategies
- Community Driven: Growing ecosystem of community-built components
- Python >= 3.11.9
- pip package manager
# Option 1: Install from PyPI (Stable)
pip install tensortrade
# Option 2: Install latest from GitHub (Development)
pip install git+https://github.com/tensortrade-org/tensortrade.git
# Option 3: Install with all dependencies for examples
pip install -r requirements.txt
pip install -r examples/requirements.txt# Run Jupyter Notebooks
make run-notebook
# Build Documentation
make run-docs
# Run Test Suite
make run-tests- User-Friendly: Designed for humans with consistent & simple APIs
- Modular: Plug-and-play components for maximum flexibility
- Extensible: Easy to add new modules and customize existing ones
- Exchanges: Connect to various trading platforms
- Feature Pipelines: Process and transform market data
- Action Schemes: Define trading actions and strategies
- Reward Schemes: Customize performance metrics
- Trading Agents: Implement learning algorithms
- Performance Reports: Track and analyze results
- Current Maintainer: Ali-hey-0
- Contact: aliheydari1381doc@gmail.com
TensorTrade is currently in Beta. While suitable for experimentation and research, use in production environments should be approached with caution.
This project is licensed under the Apache License 2.0 - see the LICENSE file for details.
Built with ❤️ by the TensorTrade community