YosuaMichael

I think in this particular case, we can do: ``` batch = torch.randint(0, 256, size=(4, 3, 224, 224)) weights = models.ViT_H_14_Weights.DEFAULT inpt = weights.transforms()(batch) model = models.vit_h_14(image_size=inpt.shape[-1]) # pass image_size...

Hi @TeodorPoncu, I think we should add the EvalPresets in https://github.com/pytorch/vision/blob/main/torchvision/transforms/_presets.py since this will be needed in the model weight , see https://github.com/pytorch/vision/blob/main/torchvision/models/optical_flow/raft.py#L538

@NicolasHug @datumbox @jdsgomes Could you check if this PR is okay to merge? I think regarding the removal of deprecated API, other than this PR only the following two things...

Just to update, I found 2 breakages internally after trying this changes. I have created diffs to fix for each of them, and I will merge this PR once the...

The failed test seems to be not related to this PR, will merge

@oke-aditya there seems lint error about the typing https://app.circleci.com/pipelines/github/pytorch/vision/21954/workflows/1179cdc7-f7db-4700-87c7-e45ef42af46c/jobs/1777451 could you resolve this?

The test failures are unrelated so we merged.

Thanks for the response! Not sure if this is relevant, but when I tried to install torchtriton using `conda install torchtriton -c pytorch-nightly` it will also install other library, here...

Oh my bad, seems like installing `torchtriton` via conda will replace the pytorch hence thats why it fixes the issue. Sorry for the misleading, it seems torchtriton is not really...