Feature: MacOs Support

[visuallization]

Hey @edbeeching

I added macos support for stable baseline 3 wrapper: python examples/stable_baselines3_example.py --env_path=./examples/JumperHard/bin/JumperHard.app --speedup=8

I haven't had time yet to look into ray, sf and clean rl yet.

[visuallization]

Okay I got this working for:

• stablebaselines 3 python examples/stable_baselines3_example.py --env_path=./examples/JumperHard/bin/JumperHard.app --speedup=8
• clearn rl python examples/clean_rl_example.py --env_path=./examples/JumperHard/bin/JumperHard.app --speedup=8

It does not yet work properly for:

• rllib gdrl --trainer=rllib --env=gdrl --env_path=./examples/JumperHard/bin/JumperHard.app --num_workers=10 --experiment=JumperHard --viz --speedup=8 --batched_sampling=True
• samplefactory gdrl --trainer=sf --env=gdrl --env_path=./examples/JumperHard/bin/JumperHard.app --num_workers=10 --experiment=JumperHard --viz --speedup=8 --batched_sampling=True

as rllib and sf do not support apple m1 gpu acceleration yet. Also see:

• https://github.com/alex-petrenko/sample-factory/issues/264
• https://github.com/ray-project/ray/issues/28321
• https://discuss.ray.io/t/rllib-pytorch-and-mac-m1-gpus-no-available-node-types-can-fulfill-resource-request/6769/9

However it is quite likely that this will already work on older intel macs, but I can't confirm because I only have a m1.

[edbeeching]

This is great! Thank you

[edbeeching]

Feel free to merge if you are happy with it