Open ealkanat opened 1 year ago
I am using Automatic 1111 and ComfyUI together. I have shared .venv folder than I deactivate it and it started to work again. Why this happened I have no idea yet :)
PYTORCH_MPS_HIGH_WATERMARK_RATIO=0.0 python3 main.py --force-fp16
Sorry guys, this error not related Ventura update. So it can be closed.
I activate Virtual Environment Folder (venv) under the Automatic 1111 installation directory and uninstalled the Torch 2.0.1 then I installed nightly version of Tourch.
pip3 uninstall torch
pip3 install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/cpu
(venv) xxxx@xxxx-MBP stable-diffusion-webui % pip3 uninstall torch
Found existing installation: torch 2.0.1
Uninstalling torch-2.0.1:
Would remove:
/Users/xxx/ai/stable-diffusion-webui/venv/bin/convert-caffe2-to-onnx
/Users/xxx/ai/stable-diffusion-webui/venv/bin/convert-onnx-to-caffe2
/Users/xxx/ai/stable-diffusion-webui/venv/bin/torchrun
/Users/xxx/ai/stable-diffusion-webui/venv/lib/python3.10/site-packages/functorch/
/Users/xxx/ai/stable-diffusion-webui/venv/lib/python3.10/site-packages/torch-2.0.1.dist-info/
/Users/xxx/ai/stable-diffusion-webui/venv/lib/python3.10/site-packages/torch/
/Users/xxx/ai/stable-diffusion-webui/venv/lib/python3.10/site-packages/torchgen/
Proceed (Y/n)? y
Successfully uninstalled torch-2.0.1
I'm having the same issue, used to use Automatic1111 but want to try ComfyUI. Getting exactly the same error but running those commands to uninstall and reinstall didn't solve the issue. Any more ideas what i can do? :)
Nevermind, it works but I had to change pip3 to pip :)
This worked, just needed to uninstall the global torch installation in my case, and reinstall the nightly.
MacBook Pro 16GB i9 AMD Radeon Pro 5500M 4 GB
missing {'cond_stage_model.clip_g.transformer.text_model.embeddings.position_ids'} 0%| | 0/20 [00:00<?, ?it/s]loc("varianceEps"("(mpsFileLoc): /AppleInternal/Library/BuildRoots/d9889869-120b-11ee-b796-7a03568b17ac/Library/Caches/com.apple.xbs/Sources/MetalPerformanceShadersGraph/mpsgraph/MetalPerformanceShadersGraph/Core/Files/MPSGraphUtilities.mm":228:0)): error: input types 'tensor<1x1024x1xf16>' and 'tensor<1xf32>' are not broadcast compatible LLVM ERROR: Failed to infer result type(s). zsh: abort PYTORCH_MPS_HIGH_WATERMARK_RATIO=0.0 python3 main.py --force-fp16 /usr/local/Cellar/python@3.10/3.10.10_1/Frameworks/Python.framework/Versions/3.10/lib/python3.10/multiprocessing/resource_tracker.py:224: UserWarning: resource_tracker: There appear to be 1 leaked semaphore objects to clean up at shutdown warnings.warn('resource_tracker: There appear to be %d '
I haven't tried before but when I used --force-fp32 I didn't get this error. But the image was rendered like this.