tensorflow/agents
A TensorFlow library providing modular, well-tested components for implementing and deploying reinforcement learning and contextual bandit algorithms.
Not currently ranked — collecting fresh signals.
star history
TF-Agents implements core RL elements as modular Agents that define Policies for environment interaction and handle training from collected experience. It provides well-tested, scalable components for fast iteration on RL algorithms, supporting implementations like DQN, PPO, and various bandit strategies. The library integrates with TensorFlow for GPU-accelerated training and includes benchmarking and testing infrastructure.
Frequently asked
- What is tensorflow/agents?
- A TensorFlow library providing modular, well-tested components for implementing and deploying reinforcement learning and contextual bandit algorithms.
- Is agents open source?
- Yes — tensorflow/agents is open source, released under the Apache-2.0 license.
- What language is agents written in?
- tensorflow/agents is primarily written in Python.
- How popular is agents?
- tensorflow/agents has 3k stars on GitHub.
- Where can I find agents?
- tensorflow/agents is on GitHub at https://github.com/tensorflow/agents.