Closed ntonic closed 2 weeks ago
Only Apple Silicon (M1/M2/M3) are supported locally on Mac.
Ok...
Imagine on Linux you support only M1 M2 M3......
On Linux it supports whatever NVIDIA/AMD support. On Mac it supports whatever Apple supports. And Apple has deliberately cut off NVIDIA cards and never implementedr AI/ML for them.
That being said, GT 750M doesn't work for this stuff on any OS, it's few generations too old.
Hello,
Krita AI diffusion won't start local server and crash........
I use Krita on MacBook Pro 2014 with NVIDIA GeForce GT 750M (2 Go dedicaded). macOS 13.6 (Ventura) with OCLP.
This graphic card run Metal 2.0 but :
2 messages in server.log :
At the beginning : INFO /Users/max/Library/Application Support/krita/ai_diffusion/server/venv/lib/python3.9/site-packages/torch/backends/mps/init.py:22: UserWarning: Skipping device NVIDIA GeForce GT 750M that does not support Metal 2.0 (Triggered internally at /Users/runner/work/pytorch/pytorch/pytorch/aten/src/ATen/mps/MPSDevice.mm:101.)
At the end : INFO raise AssertionError("Torch not compiled with CUDA enabled") 2024-09-09 03:00:40,886 INFO AssertionError: Torch not compiled with CUDA enabled 2024-09-09 03:00:41,223 ERROR