pytorch-image-models
pytorch-image-models copied to clipboard
[FEATURE] Add model weights from Bamboo
https://github.com/zhangyuanhan-ai/bamboo has released some interesting models, e.g., Bamboo-CLS ResNet-50 and Bamboo-CLS ViT B/16.
The ViT model beats most base ViTs I have seen in challenging datasets such as Objectnet.
@NightMachinery thanks, looks like it'd just be adding two sets of pretrained weights, resnet50 should be compatible already as it's torchvision compat, but the torchvision vit needs a slight remapping to timm... I'll take a closer look soon
It looks like the ViT's state_dict simply has a "module." prefix on every key. Here is a simple notebook to convert the weights: https://colab.research.google.com/drive/1s9tohf-CncnYo_y7Wdga5Tvd3bylEHOX