PyTorch: Difference between revisions

From Tygron Support wiki
Jump to navigation Jump to search
Maxim@tygron.com (talk | contribs)
No edit summary
Maxim@tygron.com (talk | contribs)
No edit summary
 
Line 3: Line 3:
PyTorch packages can be installed and managed using Conda. Conda-forge's Miniforge{{ref|Miniforge}} is an open-source application that allows you to create and manage conda environments.
PyTorch packages can be installed and managed using Conda. Conda-forge's Miniforge{{ref|Miniforge}} is an open-source application that allows you to create and manage conda environments.


Trained PyTorch Models can be exported using the [[ONNX]] file format, which in turn can be imported into the {{software}} as a [[Region-based Convolutional Neural Network (Inference Overlay)|RCNN]] for an [[Inference Overlay]].
Trained PyTorch Models can be exported using the [[ONNX]] file format, which in turn can be imported into the {{software}} as a [[Region-based Convolutional Neural Network (Inference Overlay)|Region-based Convolutional Neural Network (RCNN)]] for an [[Inference Overlay]].





Latest revision as of 15:59, 29 June 2026

PyTorch[1] is a Python-based, open source, and production-ready AI framework that supports distributed training, graph mode, and cloud platforms.

PyTorch packages can be installed and managed using Conda. Conda-forge's Miniforge[2] is an open-source application that allows you to create and manage conda environments.

Trained PyTorch Models can be exported using the ONNX file format, which in turn can be imported into the Tygron Platform as a Region-based Convolutional Neural Network (RCNN) for an Inference Overlay.