Shypd.aiThe largest open index of AI toolsbatch-ppo is an open-source reinforcement learning infrastructure project that provides efficient batched computation in TensorFlow. It extends the OpenAI Gym interface for multiple parallel environments, including an optimized implementation of Proximal Policy Optimization (PPO).
Trending
About
batch-ppo offers an optimized infrastructure for reinforcement learning, specifically designed for efficient batched computation within TensorFlow. It expands the standard OpenAI Gym interface to support multiple parallel environments, enabling agents to be developed in TensorFlow and leverage batched processing. The project includes a highly optimized implementation of Proximal Policy Optimization (PPO) as a starting point. Key components facilitate this, such as `ExternalProcess` for running environments in external processes, `BatchEnv` for handling batches of environments, and `InGraphBatchEnv` for integrating batch environments directly into the TensorFlow graph. This setup is ideal for researchers and developers looking to implement and test new reinforcement learning algorithms efficiently.
Capabilities
Pricing & Plans
Open Source
Free
FAQs
Trending
