Alexander Piskun

> > image information sent with the websocket. > > How to get the final image with python? Just open with Pillow the buffer received for message `PREVIEW_IMAGE` and that...

From the example:

The simple algorithm is like this: 1. if `bytes` arrived via websocket(add check with `isinstance` for that), check the first 4 bytes, if there is a number equal to ONE,...

I can’t imagine someone running Comfy on a CPU on macOS(PyTorch 2.3 dropped x86 macOS support as I know) Anyway the question is good, I changed code to check for...

@comfyanonymous sorry to bother you, but are you fine with this PR or not?

@comfyanonymous

The easiest way is to create an organization on github, clone all the necessary repositories there and tag them yourself after testing, imho.

no, it is not possible from technical perspective. only a separate `venv` for one **worklow** is possible

missing nodes install a higher version of `pytorch` that requires `nccl 2.19.4` Reference: https://github.com/NVIDIA/nccl/issues/1180

This is not a bug in ComfyUI, imho. And it is unrelated to hardware, **it is a software problem**. https://docs.python.org/3/library/mimetypes.html#mimetypes.init https://bugs.python.org/issue38672