M1/2/3 Apple Silicon

[Pomierny]

Hi,
I m trying to run RSS on my MacBook M1, however I ve stuck on the very beginning. I ve installed CPU-only version, however when started the script to run I get such error:

Traceback (most recent call last): File "/opt/miniconda3/envs/rss2/bin/RS2_predict", line 33, in <module> sys.exit(load_entry_point('RS2==1.0', 'console_scripts', 'RS2_predict')()) File "/opt/miniconda3/envs/rss2/lib/python3.9/site-packages/RS2-1.0-py3.9.egg/RS2/inference/predict.py", line 421, in predict_entry_point predict_from_raw_data(args.i, File "/opt/miniconda3/envs/rss2/lib/python3.9/site-packages/RS2-1.0-py3.9.egg/RS2/inference/predict.py", line 134, in predict_from_raw_data load_what_we_need(checkpoint_name) File "/opt/miniconda3/envs/rss2/lib/python3.9/site-packages/RS2-1.0-py3.9.egg/RS2/inference/predict.py", line 76, in load_what_we_need checkpoint = torch.load(checkpoint_name) File "/opt/miniconda3/envs/rss2/lib/python3.9/site-packages/torch/serialization.py", line 809, in load return _load(opened_zipfile, map_location, pickle_module, **pickle_load_args) File "/opt/miniconda3/envs/rss2/lib/python3.9/site-packages/torch/serialization.py", line 1172, in _load result = unpickler.load() File "/opt/miniconda3/envs/rss2/lib/python3.9/site-packages/torch/serialization.py", line 1142, in persistent_load typed_storage = load_tensor(dtype, nbytes, key, _maybe_decode_ascii(location)) File "/opt/miniconda3/envs/rss2/lib/python3.9/site-packages/torch/serialization.py", line 1116, in load_tensor wrap_storage=restore_location(storage, location), File "/opt/miniconda3/envs/rss2/lib/python3.9/site-packages/torch/serialization.py", line 217, in default_restore_location result = fn(storage, location) File "/opt/miniconda3/envs/rss2/lib/python3.9/site-packages/torch/serialization.py", line 182, in _cuda_deserialize device = validate_cuda_device(location) File "/opt/miniconda3/envs/rss2/lib/python3.9/site-packages/torch/serialization.py", line 166, in validate_cuda_device raise RuntimeError('Attempting to deserialize object on a CUDA ' RuntimeError: Attempting to deserialize object on a CUDA device but torch.cuda.is_available() is False. If you are running on a CPU-only machine, please use torch.load with map_location=torch.device('cpu') to map your storages to the CPU.

Not sure how can I resolve this... any suggestions?

Cheers,
Bartek

[VitoLin21]

I am sorry that I have not tested this code on a MAC.
you can try adding " -device 'cpu' " after the command
RS2_predict -i 'path/to/input' -o 'path/to/output' -m 'path/to/pretrained_model.pt'

like
RS2_predict -i 'path/to/input' -o 'path/to/output' -m 'path/to/pretrained_model.pt' -device 'cpu'

If there are same error, please modify line 76 of "Rodent-Skull-Stripping/RS2/inference/predict.py":
checkpoint = torch.load(checkpoint_name)
to:
checkpoint = torch.load(checkpoint_name, map_location=torch.device('cpu'))

Then run:
pip uninstall RS2

python setup.py install

to reinstall the framework and try RS2_predict XXX again(with -device 'cpu').

[VitoLin21]

You can also try specifying the device as -device 'mps'. I haven't tested this function, but PyTorch itself supports it (https://pytorch.org/docs/stable/notes/mps.html).