MVRL/rvsa_vitae_b

Model: ViTAE-RVSA (https://arxiv.org/abs/2208.03987)

Variant: ViTAE-b_pretrain

Example Usage:

from huggingface_hub import hf_hub_download
import torch

hf_hub_download("MVRL/rvsa_vitae_b", "model.py", local_dir=".")

from model import MaskedAutoencoderViTAE

model = MaskedAutoencoderViTAE.from_pretrained("MVRL/rvsa_vitae_b")

print(model.forward_encoder(torch.randn(1, 3, 224, 224), mask_ratio=0.0)[0].shape)