numpy errors now on new install

[colles2]

I had the setup work before, had to rebuild system other day and now get these errors...

A module that was compiled using NumPy 1.x cannot be run in
NumPy 2.0.1 as it may crash. To support both 1.x and 2.x
versions of NumPy, modules must be compiled with NumPy 2.0.
Some module may need to rebuild instead e.g. with 'pybind11>=2.12'.

...and more. I have messed around with this over past two days and all it does is cause more issues. Any ideas would be appreciated.

[bgizelarMCW]

Hi,

I just got the exact error today with NumPy. Here's roughly how I fixed it. I deleted the environment, and created a new conda yml file like so, and installed the new conda env based on this custom build file:

name: rosettafold2-1.0
channels:
  - pytorch
  - nvidia
  - defaults
  - conda-forge
  - bioconda
  - pyg
  - dglteam/label/cu121
dependencies:
  - python=3.10
  - pip
  - numpy=1.24.4               # Must stay <2.0
  - scipy=1.10.1               # Compatible with numpy 1.x
  - hhsuite=3.3.0
  - pytorch=2.2.0
  - pytorch-cuda=12.1
  - torchvision=0.17.0
  - torchaudio=2.2.0
  - dgl=2.0.0.cu121
  - pyg=2.5.0
  - pandas=2.0.3
  - requests=2.31.0
  - biopython=1.81

I then activated this new env and did the SE3Transformer portion of the install instructions. It got further, but failed to find my GPU, so inside this conda env, I did this:

conda remove pytorch torchvision torchaudio
conda install -c pytorch -c nvidia     pytorch=2.2.0     torchvision=0.17.0     torchaudio=2.2.0     pytorch-cuda=12.1

On my gpu server, with the conda env loaded, I made sure it would see my gpu:
$ python -c "import torch; print(f'CUDA available: {torch.cuda.is_available()}\nCUDA version: {torch.version.cuda}')"
CUDA available: True
CUDA version: 12.1

Hope this helps others having this same issue.

[colles2]

This fixed the issue. Thank you