timm/vit_large_patch16_224.augreg_in21k vs vit_large_patch16_224.augreg_in21k_ft_in1k'

[nighting0le01]

why is this model not avaialble in timm or torchhub??
https://huggingface.co/timm/vit_large_patch16_224.augreg_in21k

only vit_large_patch16_224.augreg_in21k_ft_in1k is visible in https://github.com/huggingface/pytorch-image-models/blob/main/timm/models/vision_transformer.py
please help!!!

[rwightman]

they are both there

[nighting0le01]

hello @rwightman thank you for your response. can you please point me to the file in the repo which has the augreg hyperparemeters and transformations you used to train

1. vit_large_patch16_224.augreg_in21k: https://huggingface.co/timm/vit_large_patch16_224.augreg_in21k
2. vit_large_patch16_224.augreg_in21k_ft_in1k: is visible in https://github.com/huggingface/pytorch-image-models/blob/main/timm/models/vision_transformer.py

I'm trying to compare vanilla distillation and finetuning on in_1k to the vit_large_patch16_224.augreg_in21k_ft_in1k which was only finetuned on 1n_1k.
Thank you again