matluster comments

Repositories
Issues
Comments

Results 24 comments of


                                            matluster

Transition to Gymnasium / Future Compatibility with Tianshou

There is coming new implementation of PolicyBasedRL in NNI v3.0. The large part of coding work is done and I just incorporate the new API of Gymnasium: https://github.com/microsoft/nni/commit/2efe0c6502d354d99181e05256823063458854ed It will...

Multimodal support?

We have people working on an example. There is currently no blocking issue.

How to run agentic rl with agent lightning in the distributed training environment with multi-nodes and multi-gpus ?

The launching should be similar to verl. Please refer to the verl example. We will need to find an environment with multi nodes and multi gpus to test this.

I find agent lightning only use 1 actor during rollout, can we launch multiple actors?

Multiple actor instances can be set via `n_workers = X` in trainer.