search

facebookresearch / BenchMARL

A collection of MARL benchmarks based on TorchRL
https://benchmarl.readthedocs.io/
MIT License
191 stars 23 forks source link

[Meltingpot] Fine tuning PPO #78

Open matteobettini opened 2 months ago

matteobettini commented 2 months ago

To execute just run meltingpot_run.py

cc @ezhang7423

ezhang7423 commented 2 months ago

I think this can probably be merged?

matteobettini commented 2 months ago

To be merged we would need https://github.com/pytorch/rl/pull/2115

Also i would like to fine tune off-policy algos as well

dziugasu commented 2 months ago

Hi @matteobettini, I'm a student doing a small research on MARL, specifically focused on meltingpot lib, though I'm kinda new to RL. Just wanted to ask if this config was tested, maybe there is a report for it?

matteobettini commented 2 months ago

Hi @matteobettini, I'm a student doing a small research on MARL, specifically focused on meltingpot lib, though I'm kinda new to RL. Just wanted to ask if this config was tested, maybe there is a report for it?

Hey, we are still working on this, but it could be a good starting point if you want to use IPPO in Meltingpot