GB7324

Graph-Break Type

Short name describing what triggered the graph break

returning intermediate with requires_grad_()

Context

Values or code snippet captured at the break point

graph output depends on source-less requires_grad_()

Explanation

Explanation of why the graph break was triggered

msg

Hints

Hints on how to resolve the graph break

• If you only need the tensor values without gradients,
• call .detach() before returning.
• Consume the gradient inside the compiled function
• (call backward() and use .grad),
• or move requires_grad_() outside torch.compile.

Additional Information

Click here to add Additional Info

Back to Registry