what algorithm do you use to sovle the overcooked problem? MADDPG?

opendilab / DI-engine

OpenDILab Decision AI Engine. The Most Comprehensive Reinforcement Learning Framework B.P.

https://di-engine-docs.readthedocs.io

Apache License 2.0

3.12k stars 377 forks source link

what algorithm do you use to sovle the overcooked problem? MADDPG? #750

Closed frandoLin closed 11 months ago

frandoLin commented 1 year ago

[ ] I have marked all applicable categories:
- [ ] exception-raising bug
- [ ] RL algorithm bug
- [ ] system worker bug
- [ ] system utils bug
- [ ] code design/refactor
- [ ] documentation request
- [ ] new feature request
[ ] I have visited the readme and doc
[ ] I have searched through the issue tracker and pr tracker

[ ] I have mentioned version numbers, operating system and environment, where applicable:

import ding, torch, sys
print(ding.__version__, torch.__version__, sys.version, sys.platform)

PaParaZz1 commented 1 year ago

We have not add the entire training pipeline for the overcooked problem in our open-source version. We tested QMIX and MAPPO in this problem. If necessary, we can add a demo in the next release version (v0.5.1).

frandoLin commented 1 year ago

We have not add the entire training pipeline for the overcooked problem in our open-source version. We tested QMIX and MAPPO in this problem. If necessary, we can add a demo in the next release version (v0.5.1).

Where can I find the training code for the overcooked problem?

PaParaZz1 commented 11 months ago

Current repo doesn't have training code for the overcooked problem, we will add a basic demo use PPO for this environment in the next version (v0.5.1 or v0.5.2).