Ask Your Question

Revision history [back]

click to hide/show revision 1
initial version

What is the recommended approach in the most recent release of PyTorch to ensure that all tensors are automatically assigned to a specific device?