Perceus

https://youtube.com/watch?v=qRACQpkDDDU

A [page](https://store.steampowered.com/app/2381500/Engine_Simulator/) has been added with `8700k` as minimum CPU - I have `6700k`. Ange [mentioned](https://github.com/Engine-Simulator/engine-sim-community-edition/issues/6#issuecomment-1500432013) cross-platform support and it is still early; the initial version released on steam will...

[Error](https://github.com/pufferai/pufferlib/compare/3.0...1d592bcae94f1a1c94ff5f840a5f213d479a1cab#files_bucket) after [modifying](https://github.com/PufferAI/PufferLib/compare/1d592bcae94f1a1c94ff5f840a5f213d479a1cab..d6670e968b26fa8502f5ed7a2962e50a39a8d613) `pufferl.py`: ``` Traceback (most recent call last): File "", line 1, in from multiprocessing.spawn import spawn_main; spawn_main(parent_pid=3204, pipe_handle=1408) ~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\Users\jayg8\AppData\Local\Programs\Python\Python313\Lib\multiprocessing\spawn.py", line 122, in spawn_main exitcode =...

Perhaps [`8bcc454`](https://github.com/pufferai/pufferlib/commit/8bcc454); not necessarily out-of-scope - may be needed.

> As for pufferlib - `bbd22d` - if starting with `device = xpu` in `pufferlib/config/ocean/target.ini`, linux shows `AssertionError: Torch not compiled with XPU enabled` which confirms the possibility. Officially without...

> Neo requires: > > [Intel(R) Graphics Compiler for OpenCL(TM)](https://github.com/intel/intel-graphics-compiler) > [Intel(R) Graphics Memory Management](https://github.com/intel/gmmlib) > Please visit their repositories for building and instalation instructions. > — [Intel](https://github.com/intel/compute-runtime/blob/master/BUILD.md) Have [yet...

 This was with stock `torch` - version 2.7.0. With Intel's torch, the familiar `RuntimeError: No XPU devices are available.` occurs. As for `puffer eval puffer_target --load-model-path latest`: `AssertionError: Torch...

Guessing not necessarily due to the recent cuda extension(s) - not `2.0`:  With `nvidia-cuda-toolkit`; changes `python setup.py clean` to: ``` No CUDA runtime is found, using CUDA_HOME='/usr' running clean...

`--vec.num-workers=4` can be used with `puffer train`. `--vec-overwork`: `demo.py` was removed in favor of the command which uses `pufferlib/pufferl.py`.

Option(s) to hide some merge(s) by some metric(s); the legend, though, right?