How to fuse the layers when using create_model

[twmht]

https://timm.fast.ai/create_model

We can use create_model to retrive the given out_indices, but it breaks the network structure into ModuleDict, I can't easily to post process the network, for example, I can't find fuse (https://github.com/huggingface/pytorch-image-models/blob/main/timm/models/repvit.py#L328) after calling create_model.

How can I do that?