[BUG] PYTORCH issue on Mac

[raeesahhy]

Hi all, Cellpose is having trouble running cyto3 on the GUI.

1. I run python -m cellpose --Zstackand upload my image. It can denoise, but does not seem able to calibrate or segment.

2. I get this error:
   NET ERROR: The operator 'aten::grid_sampler_3d' is not currently implemented for the MPS device. If you want this op to be considered for addition please comment on https://github.com/pytorch/pytorch/issues/141287 and mention use-case, that resulted in missing op as well as commit hash 2236df1770800ffea5697b11b0bb0d910b2e59e1. 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.

3. When I run 'PYTORCH_ENABLE_MPS_FALLBACK=1 python -m cellpose --Zstack' I get this error:

'NotImplementedError: Could not run 'aten::_sparse_coo_tensor_with_dims_and_tensors' with arguments from the 'SparseMPS' backend. This could be because the operator doesn't exist for this backend, or was omitted during the selective/custom build process (if using custom build). If you are a Facebook employee using PyTorch on mobile, please visit https://fburl.com/ptmfixes for possible resolutions. 'aten::_sparse_coo_tensor_with_dims_and_tensors' is only available for these backends: [MPS, Meta, SparseCPU, SparseMeta, BackendSelect, Python, FuncTorchDynamicLayerBackMode, Functionalize, Named, Conjugate, Negative, ZeroTensor, ADInplaceOrView, AutogradOther, AutogradCPU, AutogradCUDA, AutogradHIP, AutogradXLA, AutogradMPS, AutogradIPU, AutogradXPU, AutogradHPU, AutogradVE, AutogradLazy, AutogradMTIA, AutogradPrivateUse1, AutogradPrivateUse2, AutogradPrivateUse3, AutogradMeta, AutogradNestedTensor, Tracer, AutocastCPU, AutocastXPU, AutocastMPS, AutocastCUDA, FuncTorchBatched, BatchedNestedTensor, FuncTorchVmapMode, Batched, VmapMode, FuncTorchGradWrapper, PythonTLSSnapshot, FuncTorchDynamicLayerFrontMode, PreDispatch, PythonDispatcher].'

[sophiamaedler]

@raeesahhy unfortunately PyTorch has yet to implement support for sparse tensors on the MPS backend which is why you are getting this error. See some of the other issues for more info #1063 #1034

I am another fellow Mac user and for me the best solution so far has been to utilize cellpose 3.0.11. This is the most up-to-date version which still runs on the MPS backend.

[carsen-stringer]

closing for now thanks @sophiamaedler

[raeesahhy]

@sophiamaedler Thank you so very much, really appreciate it! I'll try this!