search

kcheng1021 / GaussianPro

[ICML2024] Official code for GaussianPro: 3D Gaussian Splatting with Progressive Propagation
https://kcheng1021.github.io/gaussianpro.github.io/
MIT License
551 stars 36 forks source link

A RuntimeError When I run script/waymo.sh #11

Closed duckinchina closed 3 months ago

duckinchina commented 3 months ago

Traceback (most recent call last): File "/data/GaussianPro/train.py", line 372, in 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 "/data/GaussianPro/train.py", line 214, in training loss = (1.0 - opt.lambda_dssim) Ll1 + opt.lambda_dssim (1.0 - ssim(image, gt_image, mask=opacity_mask)) File "/data/GaussianPro/utils/loss_utils.py", line 41, in ssim return _ssim(img1, img2, window, window_size, channel, size_average, mask=mask) File "/data/GaussianPro/utils/loss_utils.py", line 44, in _ssim mu1 = F.conv2d(img1, window, padding=window_size // 2, groups=channel) RuntimeError: Expected 4-dimensional input for 4-dimensional weight [3, 1, 11, 11], but got 3-dimensional input of size [3, 1067, 1600] instead

I don't know why this happened. Could you help me? Thanks!

duckinchina commented 3 months ago

The authors have done a great work. I don't know whether it's just because the code is still refreshing, because I haven't found train_baseline.sh and train.sh in the scripts folder. Thank you.

kcheng1021 commented 3 months ago

Thanks for your acknowledgement.

  1. What is your pytorch version? We have the same requirement for the PyTorch version as 3DGS. We are using PyTorch 1.12.1.
  2. Thanks for your reminder. We split the scripts by the scenes. The parameters for baseline and gaussianpro can be found in either waymo.sh or demo.sh.
duckinchina commented 3 months ago

Hi, this problem has been solved by changing the cuda version from 11.1 to 11.7. Many thanks for the authors' advice!