Ross Wightman

@pfeatherstone actually had a use case for doing this for 3d come up recently, think it might be a need extension, would make sense to cover 1d as well if...

@pfeatherstone thanks for sharing that snippet, I think initial approach would be to try as a helper that mutates an existing model like above, adding arguments through the create_model fn...

@mywebinfo65536 Some of the maxvit models are currently resizable, any of the models with rmlp in the name will resize if you pass `img_size` the model on creation, BUT it...

@xvr-hlt yeah, if someone wants to tackle the interpolation code (a bit different for both the TF style and swin style) that'd be great. There's some code to base this...

maxvit, coatnet, and swin models support resize on load now, it would be possible to support dynamic resize as well but would have a small runtime penalty and more complexity...

@MohamedAliRashad you need a version of timm with that model, recent 0.8.x pre-releases / main branch have it.

Fixed in https://github.com/rwightman/pytorch-image-models/commit/2cfff0581b643347673cd19713cb3ea3b09c77a9

@xwang233 PR welcome, there is a significant shift in both docs and models right now, 0.6.x (pypi releases or this branch https://github.com/rwightman/pytorch-image-models/tree/0.6.x should be used for stability). On `main`, that...

I'm slowly making my way through every model on main (0.8.x) and converting them to this new `architecture.pretrained_tag` form, I do intend to add a backwards compatible deprecation mapping, but...

@xwang233 @kaczmarj deprecation mapping feature on main branch now, will be an updated 0.8 pre-release within a day or two with that. Still have some models to add, but efficientnet...