xlnstorch.benchmark.Benchmark.post_forward

Benchmark.post_forward(output)

Optional light-weight processing of the forward result. Called outside the timing region so it has no effect on latencies.

Parameters:

output (Any)

Return type:

None