What features do you want to add to XingTian?

[tmhm]

We want to know what you want to add to XingTian, we will evaluate your suggestions and make a plan. Please let us know if you are interested.

• [ ] add detailed user-defined module guidance
• [x] add supports for distribution of continuous & discrete action space
• [x] add supports for model-based algorithms, e.g, MuZero
• [x] add supports for multi-agents algorithms, e.g, Qmix
• [ ] add support for evolutionary algorithms, e.g, PBT
• [x] add supports for DaVinci
• [ ] add supports for training with Multi-GPU
• [ ] add supports for call XingTian within user-python-code

[terryzhao127]

Firstly, XingTian is a great distributed reinforcement learning framework with elegant codes and clear modularization. Following are the features that can be added:

1. More docs on the zeus package for developers interested in the implementation
2. K8s support

[tmhm]

Thanks for your advices. The docs will be enriching consciously , and support for K8s is necessary. By the way, if you are interested in them, we'd be happy to receive your contributions. :)