You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: examples/profiling/README.md
+5-1Lines changed: 5 additions & 1 deletion
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -339,4 +339,8 @@ there are still some gaps outside the compiled regions. A full compilation
339
339
will likely mitigate it. In case it doesn't, the above observations could
340
340
be useful to mitigate that.
341
341
* Use of CUDA Graphs can also help mitigate CPU overhead related issues. CUDA Graphs can be enabled by setting the `torch.compile` mode to `"reduce-overhead"` or `"max-autotune"`.
342
-
* Diffusers' integration of `torch.compile` is documented [here](https://huggingface.co/docs/diffusers/main/en/optimization/fp16#torchcompile).
342
+
* Diffusers' integration of `torch.compile` is documented [here](https://huggingface.co/docs/diffusers/main/en/optimization/fp16#torchcompile).
343
+
344
+
## Acknowledgements
345
+
346
+
Thanks to [vkuzo](https://github.com/vkuzo) and [jbschlosser](https://github.com/jbschlosser) from the PyTorch team for providing invaluable feedback on the guide.
0 commit comments