GB9158

Graph-Break Type

Short name describing what triggered the graph break

Can’t convert torch._check*() message closure

Values or code snippet captured at the break point

str(message_vt)

Explanation of why the graph break was triggered

The message argument of torch._check*() must be a function whose closure variables are Python constants.

Hints on how to resolve the graph break

Remove closure variables that reference non-constant values, e.g.
remove references to tensor x in lambda: f'{x} failed check'
It may be possible to write Dynamo tracing rules for this code. Please report an issue to PyTorch if you encounter this graph break often and it is causing performance issues.