This is a forked version of https://github.com/openai/baselines. The forking
is dated March 21st, 2019; with last commit set to 1b09243.
The purpose of this forked version is to implement Learning to reinforcement learn by Wang et al., since no official implementation is provided by DeepMind. Our implementation is partially dependent on the NASGym.
In the paper, a deep meta-reinforcement learning algorithm is implemented using the A2C setting, therefore we perform some modification on top of the A2C implementation on this baselines to come up with the meta version.
We hope that this code can be generalized to different environments that implement the OpenAI Gym Environment interface too.
You can run the meta-A2C as any other baselines, more information here.