[bug]: Strange behaviour on Mac OS

[GitKernelDesign]

Is there an existing issue for this problem?

• I have searched the existing issues

Operating system

macOS

GPU vendor

Apple Silicon (MPS)

GPU model

M3 Pro

GPU VRAM

18Go

Version number

5.5.0

Browser

Electron Application

Python dependencies

No response

What happened

-Bug 1-
I have this error (with application):

NotImplementedError: The operator 'aten::__rshift__.Tensor' is not currently implemented for the MPS device. If you want this op to be added in priority during the prototype phase of this feature, please comment on https://github.com/pytorch/pytorch/issues/77764. As a temporary fix, you can set the environment variable PYTORCH_ENABLE_MPS_FALLBACK=1 to use the CPU as a fallback for this op. WARNING: this will be slower than running natively on MPS.

Fixing PYTORCH_ENABLE_MPS_FALLBACK=1 in Venv/bin/activate does not work

BUT

No problem running via the Terminal

cd ~/invokeai
source .venv/bin/activate
invokeai-web

with or without fixing PYTORCH_ENABLE_MPS_FALLBACK=1 everything works.

---

-Bug 2-

Patchmatch does not compile via the application. You have to use the Terminal as before.

What you expected to happen

I'd like it to work in the application.

How to reproduce the problem

No response

Additional context

No response

Discord username

No response

[dhlsam]

same on m4 mac