Yineng Xiong

I think in RandomFlip3D ```Python input_dict['cam2img'][0][2] = w - input_dict['cam2img'][0][2] ``` should be ```Python input_dict['cam2img'][0][2] = w - input_dict['cam2img'][0][2] input_dict['cam2img'][0][3] = -input_dict['cam2img'][0][3] ``` because in KITTI dataset, `P2[:, 3]` is...

Any schedule to support image enhancement / photo retouching tasks?

MiDaS v2.1 small model missing

1

the link to MiDaS v2.1 small model is missing

Can you provide other nFrames pretrained model like nFrames=3? thanks

If i wanna froze the checkpoint to a .pb file which node should i use?

odometry_poses file

5

It seems the odometry_poses file downloaded from KITTI website is different from the poses file provided in this repo

Train Stucked

4

Hi ~ I Use ` if torch.cuda.device_count() > 1: model = torch.nn.DataParallel(model) model = bnconvert(model) model.cuda() ` to use sync-bn during multi-gpu training, but when training the network, it looks...

Great work! I have a question about training on Adobe FiveK dataset. Do you crop 100 * 100 patches during training or just use the original size with batchsize=1?