dhlab-epfl / dhsegment-torch Goto Github PK

We are trying to run the dhSegment-torch on Mac M1 Gpu. However, getting the following error

"NotImplementedError: convolution_overrideable not implemented. You are likely triggering this with tensor backend other than CPU/CUDA/MKLDNN, if this is intended, please use TORCH_LIBRARY_IMPL to override this function
"

Could you please help in this regard? Also, how to set the device as "mps" for running the code on Mac M1?

The random seed initialisation using numpy might be faulty:

https://www.reddit.com/r/MachineLearning/comments/mocpgj/p_using_pytorch_numpy_a_bug_that_plagues/

Many thanks for checking and, ideally, correcting. Happy to help if needed.

Hi y'all. If I try to use the environment.yml in Ubuntu 22.04 as described, I get a dependency error:

`LibMambaUnsatisfiableError: Encountered problems while solving:

• nothing provides requested numpy 1.21.0**
• nothing provides requested tensorboard 2.4.1**
• package pandas-1.3.2-py37h8c16a72_0 requires python >=3.7,<3.8.0a0, but none of the providers can be installed

Could not solve for environment specs
The following packages are incompatible
├─ numpy 1.21.0** does not exist (perhaps a typo or a missing channel);
├─ pandas 1.3.2** is installable with the potential options
│ ├─ pandas 1.3.2 would require
│ │ └─ python >=3.7,<3.8.0a0 , which can be installed;
│ ├─ pandas 1.3.2 would require
│ │ └─ python >=3.8,<3.9.0a0 , which can be installed;
│ └─ pandas 1.3.2 would require
│ └─ python >=3.9,<3.10.0a0 , which can be installed;
├─ python 3.10.11** is not installable because it conflicts with any installable versions previously reported;
└─ tensorboard 2.4.1** does not exist (perhaps a typo or a missing channel).`

If I try Anaconda in Windows instead, I get a similar error:
`ResolvePackageNotFound:

• numpy=1.21.0
• tensorboard=2.4.1`

Hi, I was going through the demo was wondering whether structure for the Ground Truth could be documented to perform training on a custom dataset from scratch ?