wuqun-tju

visualization

11

hello， Could you provide visuliaztion code generating Fig6 in the paper

Thank you for your excellent work! I am very interested in it . Is it possible to provide inference demo and pre-trained model. I wanna test the final extrinsic in...

Hello, Thank you for your excellent work! I want to get cameras poses using your model, can you provide a inference demo!

About training command and train_loss, val_loss

7

Thank you for your excellent work! I have some questions about training. 1. In the train command line you used for training your models, there isn't dataset 2. About training...

About 3 steps in training

1

Thanks for your excellent work! In the process of training, there are 3 steps, the head of first two is linear, and the last one is dpt. I wonder whether...

About center point when training

1

In the training step, I found there is a step cropping image according camera_intrinsic and make principal point to be center of image in https://github.com/naver/dust3r/blob/main/dust3r/datasets/co3d.py#L100 . I wonder cropping is...

Failure on training

2

Hello ， when I try to train the model using blendedmvs, it core in losses.py assert valid_matches.shape == torch.Size([B, N]) and valid_matches.sum() > 0 can you help me solve it,...

About test_criterion

1

Hello, I wanna know why the sign of MatchingLoss in the test_criterion is negtivate: Regr3D(L21, norm_mode='?avg_dis', gt_scale=True, sky_loss_value=0) + -1.*MatchingLoss(APLoss(nq='torch', fp=torch.float16), negatives_padding=12288)"

When I run demo, it reported this error  How to solve it

About inference intrinsics

1

Hello, I try to infer intrinsics, I used gt_intrinsic to initiate the camera, and check the output["intrinsics"] of the model, but pred_intrinsics fx is twice than gt_intrinsic, eg: fx_pred =...