Joseph Suarez

Hello -- This repo is archived. Please see jsuarez5341/neural-mmo for the latest version and raise issues there if you feel the current build is still unclear.

Apologies for not seeing this earlier -- this is no longer the main repo. Please raise issues on my fork, jsuarez5341/neural-mmo, in the future. It seems you are attempting to...

Hello -- This repo is archived. Please see jsuarez5341/neural-mmo for the latest version and raise issues there if you feel the current build is still unclear. For reference, map files...

Late but needed, as this post is still high up on google. The reason this is happening is that render(scene, camera) converts the given scene into kivy instructions and adds...

This really needs a fix. Gross behavior on an otherwise great library

This looks reasonable, waiting for hardware to test end to end. Any other optimization ideas for split? It's the main bottleneck right now. From before your patch: 🐡 python tests/test_extensions.py...

The start of this thread is from 0.4, super old version As of 1.0, I have it running with Python 3.11 by default. I don't think I have tested 3.12....

I can almost guarantee that I can bring the RAM usage way down... porting it now, will need a bit of time. CPU cores will still be non-negotiable, the env...

Found it! https://github.com/PWhiddy/PokemonRedExperiments/blob/fa01143e4b8d165136199be7155757495d16e56a/baselines/run_baseline_parallel.py#L62 n_steps is the number of frames per environment that you are keeping in memory. So 2048*8 for each of 44 environments... 720,896. Napkin math says 44 GB...

Okay, I got it training with 8GB of RAM. Will see if it learns anything