Can't setup the environment

[dmarcos]

 > conda env create -f environment.yml

Collecting package metadata (repodata.json): | WARNING conda.models.version:get_matcher(535): Using .* with relational operator is superfluous and deprecated and will be removed in a future version of conda. Your spec was 1.8.0.*, but conda is ignoring the .* and treating it as 1.8.0
WARNING conda.models.version:get_matcher(535): Using .* with relational operator is superfluous and deprecated and will be removed in a future version of conda. Your spec was 1.9.0.*, but conda is ignoring the .* and treating it as 1.9.0
done
Solving environment: failed

ResolvePackageNotFound:
  - torchvision==0.11.1=py39_cu113
  - cudatoolkit==11.3.1=h2bc3f7f_2
  - pytorch==1.10.0=py3.9_cuda11.3_cudnn8.2.0_0
  - torchaudio==0.10.0=py39_cu113

After following the instructions. Then:

   > conda activate prisma

   Could not find conda environment: prisma

Just followed the instructions verbatim. Don't know enough about the pytorch stack to troubleshoot myself. Thanks!

[patriciogonzalezvivo]

Do you mind sharing what OS and conda version are you using?

[dmarcos]

MacOS 13.4.1
conda 4.12.0

[patriciogonzalezvivo]

It makes more sense, those errors are because CUDA can't be installed ( notice all missing packages refers to cu113 == CUDA 11.3 ). I will add in the roadmap to explore support for MacOS

[dmarcos]

Thanks! I guess also requires an Nvidia GPU. Might be good to mention on the README

[patriciogonzalezvivo]

To be clear, in theory it's possible to use Metal backend instead of CUDA. https://pytorch.org/blog/introducing-accelerated-pytorch-training-on-mac/
There are some changes that need to be make, I just don't have experience using metal.

Would be great if someone can help with that.