Issue about using dinov2 when generating 2d embedding

[LydJason]

When I tried to run project_feature.py, I encountered problem at :

Part2Object/pseudo_mask_gen/project_feature.py

Line 183 in 65d728e

image_embedding = dinov2_vitb14(color).reshape((68, 91, 768)).permute(2, 0, 1).unsqueeze(0)

I print the shape of dinov2_vitb14(color), it seem to be [1, 768], which can't be reshape to [68, 91, 768]:

$ python pseudo_mask_gen/project_feature.py
projecting multiview features to point cloud...
processing scene0000_00...
torch.Size([1, 768])
Traceback (most recent call last):
  File "/home/lyd/Part2Object/pseudo_mask_gen/project_feature.py", line 185, in <module>
    image_embedding = image_embedding.reshape((68, 91, 768)).permute(2, 0, 1).unsqueeze(0)
RuntimeError: shape '[68, 91, 768]' is invalid for input of size 768

Here, I use the full version of scannet, for each color image come with 1296*968 pixels, the dino model I use is dinov2_vitb14_pretrain.pth according to the original code.
Am I using the wrong input? Please let me know, thank you.

[zhangyl4]

This error is strange and we can't seem to reproduce your problem. Can you get features for any selected test image? Or provide more detailed information about how the program runs?

[liuxuexun]

@LydJason Do you solve it? I meet the same problem, thank you!

[liuxuexun]

This error is strange and we can't seem to reproduce your problem. Can you get features for any selected test image? Or provide more detailed information about how the program runs?

The shape of the tensor color is torch.Size([1, 3, 952, 1274]), and when I run image_embedding = dinov2_vitb14(color).reshape((68, 91, 768)).permute(2, 0, 1).unsqueeze(0) in here meet the same error as LydJason.

Is it the error cause by dinov2?

Is the input tensor color in your reproduction also be torch.Size([1, 3, 952, 1274])?

[liuxuexun]

I have read the code of dinov2, the output of dinov2_vitb14(color) may be in here. It's name is clstoken, I guess it is the summary of the whole picture.