same architecture with timm/vit_base_patch14_dinov2.lvd142m

git clone https://github.com/DepthAnything/Depth-Anything-V2
cd Depth-Anything-V2

translate

'''
wget https://huggingface.co/depth-anything/Depth-Anything-V2-Small/resolve/main/depth_anything_v2_vits.pth?download=true
wget https://huggingface.co/depth-anything/Depth-Anything-V2-Base/resolve/main/depth_anything_v2_vitb.pth?download=true
wget https://huggingface.co/depth-anything/Depth-Anything-V2-Large/resolve/main/depth_anything_v2_vitl.pth?download=true
'''
import torch

from depth_anything_v2.dpt import DepthAnythingV2

DEVICE = 'cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu'

model_configs = {
    'vits': {'encoder': 'vits', 'features': 64, 'out_channels': [48, 96, 192, 384]},
    'vitb': {'encoder': 'vitb', 'features': 128, 'out_channels': [96, 192, 384, 768]},
    'vitl': {'encoder': 'vitl', 'features': 256, 'out_channels': [256, 512, 1024, 1024]},
}

encoder = 'vitb' # or 'vits', 'vitb'

model = DepthAnythingV2(**model_configs[encoder])
model.load_state_dict(torch.load(f'depth_anything_v2_{encoder}.pth?download=true', map_location='cpu'))
vit = model.pretrained

# total_params = 0
# for name, param in vit.named_parameters():
#     print(f"Parameter: {name} - Size: {param.size()} - Total Elements: {param.numel()}")
#     total_params += param.numel()
# print(f"Total number of parameters in ViT: {total_params}")

filtered_state_dict = {k: v for k, v in vit.state_dict().items() if 'mask_token' not in k}
torch.save(filtered_state_dict, "pytorch_model.bin")

usage

from urllib.request import urlopen
from PIL import Image
import timm

img = Image.open(urlopen(
    'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png'
))

model = timm.create_model(
    'vit_base_patch14_dinov2.lvd142m',
    pretrained=True,
    num_classes=0,  # remove classifier nn.Linear
    checkpoint_path="pytorch_model.bin"
)

# model2.load_state_dict(torch.load("backbone_weights.pth"))

# for name, param in model.named_parameters():
#     print(f"Parameter: {name} - Size: {param.size()} - Total Elements: {param.numel()}")
model = model.eval()

# get model specific transforms (normalization, resize)
data_config = timm.data.resolve_model_data_config(model)
transforms = timm.data.create_transform(**data_config, is_training=False)

output = model(transforms(img).unsqueeze(0))  # output is (batch_size, num_features) shaped tensor

# or equivalently (without needing to set num_classes=0)

output = model.forward_features(transforms(img).unsqueeze(0))
# output is unpooled, a (1, 1374, 1024) shaped tensor

output = model.forward_head(output, pre_logits=True)
print(output)