pytorch-directml 1.8.0a0.dev211021


Tensors and Dynamic neural networks in Python with strong GPU acceleration

Tensors and Dynamic neural networks in Python with strong GPU acceleration

Stars: 817, Watchers: 817, Forks: 120, Open Issues: 51

The microsoft/DirectML repo was created 2 years ago and was last updated 10 hours ago.
The project is popular with 817 github stars!

How to Install pytorch-directml

You can install pytorch-directml using pip

pip install pytorch-directml

or add it to a project with poetry

poetry add pytorch-directml

Package Details

Microsoft Corporation
GitHub Repo


  • Scientific/Engineering
  • Scientific/Engineering/Artificial Intelligence
  • Scientific/Engineering/Mathematics
  • Software Development
  • Software Development/Libraries
  • Software Development/Libraries/Python Modules
No  pytorch-directml  pypi packages just yet.


A list of common pytorch-directml errors.

No  pytorch-directml  errors just yet.

Code Examples

Here are some pytorch-directml code examples and snippets.

No  pytorch-directml  code examples just yet.

GitHub Issues

The pytorch-directml package has 51 open issues on GitHub

  • resnet50 tutorial failed for me
  • argmax function cause error (pytorch-directml)
  • Update component governance pipeline with updated build pool
  • Append and shift elements operator?
  • Roll GPGMM and track SV heaps used to clear.
  • .to('dml') not working
  • Could not run 'aten::index_select' on RX 6700
  • onnxruntime-directml slower than onnxruntime cpu
  • Bump opencv-python from to in /TensorFlow/yolov3
  • Bump pillow from 7.0.0 to 8.3.2 in /TensorFlow/squeezenet/src
  • DirectML is x2.8 slower than CUDA

See more issues on GitHub