Open Yuhuoo opened 2 months ago
But if I set sky_seg to True, the program doesn't run properly.
Are all the images in your dataset of the same resolution?
When I run the train script with the cfg_args as follows:
Namespace(data_device='cuda', eval=True, images='images', load_depth=False, load_normal=False, model_path='output/tea_room874', resolution=-1, sh_degree=3, sky_seg=False, source_path='/data1/GaoAo/code/GaussianPro/dataset/tea_room_874/OPENCV', white_background=False)
I have got the error:
Traceback (most recent call last): File "train.py", line 372, in <module> training(lp.extract(args), op.extract(args), pp.extract(args), args.test_iterations, args.save_iterations, args.checkpoint_iterations, args.start_checkpoint, args.debug_from) File "train.py", line 172, in training src_K.unsqueeze(0), src_pose.unsqueeze(0), thre1=2, thre2=0.01) File "/data1/GaoAo/code/GaussianPro/utils/graphics_utils.py", line 344, in check_geometric_consistency outputs = reproject_with_depth(*inputs) File "/data1/GaoAo/code/GaussianPro/utils/graphics_utils.py", line 319, in reproject_with_depth sampled_depth_src = bilinear_sampler(depth_src.view(batch, 1, height, width), torch.stack((x_src, y_src), dim=-1).view(batch, height, width, 2)) RuntimeError: shape '[1, 1, 1195, 1600]' is invalid for input of size 1707200
Is this because I set the --sky_seg as false ?
Hi, did you solve the problem? I met the same error. Thanks
When I run the train script with the cfg_args as follows:
Namespace(data_device='cuda', eval=True, images='images', load_depth=False, load_normal=False, model_path='output/tea_room874', resolution=-1, sh_degree=3, sky_seg=False, source_path='/data1/GaoAo/code/GaussianPro/dataset/tea_room_874/OPENCV', white_background=False)
I have got the error:
Is this because I set the --sky_seg as false ?