Hi @weiglt we’re currently developing on MacOS, primarily Apple Silicon.
Currently, there are some dependencies on our project that aren’t compatible with Windows but the team is talking about this and we may have more info later.
I wrote up a few additional details on Windows OS in this GitHub Issue. In summary, Monty is independent of Habitat (which does not run on Windows). However, we have some unintended coupling to Habitat that needs to be removed before Monty will run on Windows as expected.
hi @cortex we do not have a timeline currently, but this improvement is high up in our list of things to do. See this post for more info about the proposed approach: The Next Monty 3D Environment Simulator
Hi @cortex, in addition to what @brainwaves mentioned, a short-term obstacle is unintentional coupling with HabitatSim. I removed one source of coupling but discovered another as described in this GitHub comment. I haven’t seen anyone comment that they are working on removing the remaining coupling so that obstacle remains in place. I’m working on other things at the moment, but if I return to the issue, I’ll comment, letting folks know that I’m looking at it again.
The remaining unintended coupling to HabitatSim during benchmark or experiment run is being removed in the scope of pull request #153. Once that pull request is merged (or if you want to use the pull request branch), running a (non-HabitatSim) benchmark or experiment without HabitatSim installed should work, i.e., python benchmarks/run.py -e hand_intrusion_world_image_on_scanned_model.
Tests continue to depend on HabitatSim, so pytest will still complain.