Some questions about evaluation.

[111222444555]

After reproduction, the evaluation code failed to yield the experimental results presented in Table 1 of the paper.The project code, model and evaluation code used are all available at https://github.com/UVA-Computer-Vision-Lab/ovmono3d.

[yaojin17]

Hi, the existing code is for the v1 version of the model. In the updated paper, the results are from the updated model (with an additional depth module). I will update this codebase to the v2 model in the near future.

[111222444555]

Thank you very much.

[BozhouZha]

Hi, the existing code is for the v1 version of the model. In the updated paper, the results are from the updated model (with an additional depth module). I will update this codebase to the v2 model in the near future.

Hi there, I found the released weights checkpoint of ovmono3d-lift has no depth estimator, instead using the cube head to predict the depth z (in camera frame). Is this the v1 version of ovmono3d-lift? And will there be any update regarding the v2 model? Thanks!