The size of tensor a(162) must match the size of tensor b(3) at non-singleton dimension 1
Opened this issue · 4 comments
Hello, author! I have a question to ask you. When I run the code, I report the following errors. What is the reason? I didn't change the code.
What's wrong? I cannot see the uploaded image.
What's wrong? I cannot see the uploaded image.
The wrong with "The size of tensor a(162) must match the size of tensor b(3) at non-singleton dimension 1"
File "/home/ubuntu/ygx/1/NerfingMVS-main/src/run_nerf.py", line 137, in render
all_ret['depth_map'] = -1/rays_d_ori[:, 2]*(1 / (1 - all_ret['depth_map']) + rays_o_ori[:, 2])
File "/home/ubuntu/ygx/1/NerfingMVS-main/src/run_nerf.py", line 162, in render_path
rgb, disp, acc, depth, _ = render(H, W, focal, depth_priors=depth_priors[i], depth_confidences=depth_confidences[i], chunk=chunk, c2w=c2w[:3,:4], **render_kwargs)
File "/home/ubuntu/ygx/1/NerfingMVS-main/src/run_nerf.py", line 624, in train
image_list=image_list
File "/home/ubuntu/ygx/1/NerfingMVS-main/run.py", line 22, in
run_nerf.train(args)
Since I use '--no_ndc' for all scannet scenes, I do not test this line. In my opinion, the depth map in Cartesian coordinate is not equal to that in ndc coordinate.
Since I use '--no_ndc' for all scannet scenes, I do not test this line. In my opinion, the depth map in Cartesian coordinate is not equal to that in ndc coordinate
OK, thank you very much for your reply. I'll try ‘--no_ndc’ this situation.