Summary

Sourcehttps://pytorch.org/

License:  https://github.com/pytorch/pytorch/blob/master/LICENSE

Path:  part of conda installations

Documentationhttps://pytorch.org/

PyTorch is a Python package that provides two high-level features:

  • Tensor computation (like NumPy) with strong GPU acceleration
  • Deep neural networks built on a tape-based autograd system
You can reuse your favorite Python packages such as NumPy, SciPy and Cython to extend PyTorch when needed.

Using  pytorch

To initialize the environment use the module command for conda/3.6:

[elx]$ module load maxwell conda/3.6
# use the "standard pytorch
[elx]$ conda list | grep -i pytorch
pytorch                   1.0.1           py3.6_cuda9.0.176_cudnn7.4.2_0    pytorch
[elx]$ python
Python 3.6.7 | packaged by conda-forge | (default, Feb 28 2019, 09:07:38) 
[GCC 7.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import torch
>>> torch.__version__
'1.0.1'

# use the pytorch conda environment
[elx]$ conda env list | grep -i pytorch
pytorch                  /software/anaconda3/5.2/envs/pytorch
[elx]$ python
Python 3.6.7 | packaged by conda-forge | (default, Feb 28 2019, 09:07:38) 
[GCC 7.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import torch
>>> torch.__version__
'1.0.1'
>>> import torchvision
>>> torchvision.__version__
'0.2.1'
[elx]% conda deactivate

# get some information
[elx]% xwhich pytorch

 No executable named pytorch found 
 Documentation: https://confluence.desy.de/display/IS/pytorch
 URL:           https://pytorch.org/
 Manual:        https://pytorch.org/
 License:       https://github.com/pytorch/pytorch/blob/master/LICENSE