What RL library supports custom LSTM and Transformer neural networks to use with algorithms such as PPO?

• popgym

  4 143 6.1 Python

  Partially Observable Process Gym

  

POPGym is based on RLlib and has two linear transformers and five or six RNN variants, including LSTM. I've found that transformers tend to perform pretty poorly in RL when compared to RNNs.

• ppo-implementation-details

  18 557 0.0 Python

  The source code for the blog post The 37 Implementation Details of Proximal Policy Optimization

I am still working on it, but I used the ppo implementation of https://github.com/vwxyzjn/ppo-implementation-details and modifiy it. Fir transformer, i just implement with pytorch.

• InfluxDB

  www.influxdata.com featured

  Power Real-Time Data Analytics at Scale. Get real-time insights from all types of time series data with InfluxDB. Ingest, query, and analyze billions of data points in real-time with unbounded cardinality.

  

• episodic-transformer-memory-ppo

  5 109 2.5 Python

  Clean baseline implementation of PPO using an episodic TransformerXL memory

I provide baseline implementations on TransformerXL + PPO and LSTM/GRU + PPO. These are designed to be slim and easy-to-follow so that you can advance those implementations to the features and toolset that you need.

• recurrent-ppo-truncated-bptt

  6 105 5.1 Jupyter Notebook

  Baseline implementation of recurrent PPO using truncated BPTT

I provide baseline implementations on TransformerXL + PPO and LSTM/GRU + PPO. These are designed to be slim and easy-to-follow so that you can advance those implementations to the features and toolset that you need.

Suggest a related project