| FazBrowse GitHub Viewer | Trending | | Home |
| Tools: [Download Repo ZIP] [Original HTTPS Page] |
There was a problem hiding this comment.
Looks good 👍🏽 Thanks. Minor comments.
Sorry, something went wrong.
| expected_set = {"HunyuanVideoTransformer3DModel"} | ||
| super().test_gradient_checkpointing_is_applied(expected_set=expected_set) | ||
|
|
||
| @require_torch_gpu |
There was a problem hiding this comment.
Would make sense to add these decorators to the top of the Mixin no?
Sorry, something went wrong.
There was a problem hiding this comment.
These decorators are present in the mixin:
diffusers/tests/models/test_modeling_common.py
Lines 1762 to 1765 in b5c2050
Sorry, something went wrong.
|
|
||
| container: | ||
| image: diffusers/diffusers-pytorch-compile-cuda | ||
| image: diffusers/diffusers-pytorch-cuda |
There was a problem hiding this comment.
Small nit. We can remove the -k "compile" in the test runner step.
Sorry, something went wrong.
|
Thanks for the reviews, @DN6! Hope to make our compile CI better and better. |
Sorry, something went wrong.
| Back | FazBrowse Home | New Git URL |
What does this PR do?
Regarding the final point, I think we already test torch.compile() support for popular models and doing it at the pipeline-level makes little sense to me. But no strong opinions.