[BUG] AttributeError: 'RotaryEmbeddingDinoV3' object has no attribute 'update_feat_shape'

[ananthu-aniraj]

Hi, thanks for the great library.

I was trying to use the DinoV3 models and faced the following error when using the set_input_size() function :

AttributeError: 'RotaryEmbeddingDinoV3' object has no attribute 'update_feat_shape'

To Reproduce

import timm
model = timm.create_model('vit_base_patch16_dinov3.lvd1689m', pretrained=True)
model.set_input_size(512)

Possible Fix
Add the update_feat_shape() function to the RotaryEmbeddingDinoV3 class.

[rwightman]

Curious why you are calling set input size there, dynamic image size is enabled by default for dinov3 so it will accept 512 images without the call. But I should fix that for a no-op regardless

…

On Fri, Oct 3, 2025, 4:19 AM Ananthu Aniraj ***@***.***> wrote: Assigned #2593 <#2593> to @rwightman <https://github.com/rwightman>. — Reply to this email directly, view it on GitHub <#2593 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABLQICDWNQDCK323TXLSEKT3VZLTRAVCNFSM6AAAAACIGJP2UKVHI2DSMVQWIX3LMV45UABCJFZXG5LFIV3GK3TUJZXXI2LGNFRWC5DJN5XDWMRQGA3TSMZQHE2DMNY> . You are receiving this because you were assigned.Message ID: <huggingface/pytorch-image-models/issue/2593/issue_event/20079309467@ github.com>