RuntimeError: Expected 4-dimensional input for 4-dimensional weight [3, 1, 11, 11], but got 3-dimensional input of size [3, 900, 1600] instead

[yuxuanfanOrion]

Hi! Thanks for your great work.

I have met some bugs when I try to train bungee dataset.

The following are the bugs. Thanks for helping!

Traceback (most recent call last): File "/home/fanyx2311/.conda/envs/sgs/lib/python3.8/runpy.py", line 194, in _run_module_as_main return _run_code(code, main_globals, None, File "/home/fanyx2311/.conda/envs/sgs/lib/python3.8/runpy.py", line 87, in _run_code exec(code, run_globals) File "/home/fanyx2311/.local/lib/python3.8/site-packages/debugpy/main.py", line 39, in cli.main() File "/home/fanyx2311/.local/lib/python3.8/site-packages/debugpy/server/cli.py", line 430, in main run() File "/home/fanyx2311/.local/lib/python3.8/site-packages/debugpy/server/cli.py", line 284, in run_file runpy.run_path(target, run_name="main") File "/home/fanyx2311/.local/lib/python3.8/site-packages/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py", line 321, in run_path return _run_module_code(code, init_globals, run_name, File "/home/fanyx2311/.local/lib/python3.8/site-packages/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py", line 135, in _run_module_code _run_code(code, mod_globals, init_globals, File "/home/fanyx2311/.local/lib/python3.8/site-packages/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py", line 124, in _run_code exec(code, run_globals) File "train.py", line 527, in training(lp.extract(args), op.extract(args), pp.extract(args), dataset, args.test_iterations, args.save_iterations, args.checkpoint_iterations, args.start_checkpoint, args.debug_from, wandb, logger) File "train.py", line 143, in training ssim_loss = (1.0 - ssim(image, gt_image)) File "/data20_ssd/fanyx2311/Scaffold-GS/utils/loss_utils.py", line 41, in ssim return _ssim(img1, img2, window, window_size, channel, size_average) File "/data20_ssd/fanyx2311/Scaffold-GS/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, 900, 1600] instead Training progress: 0%|

[farewellwho]

I've got the same exception. Did you solve the problem?

[farewellwho]

I've got the same exception. Did you solve the problem?

This problem happened because I was using the wrong versions. My cuda toolkit version was 11.1, and torch related version were torch==1.10.2+cu111, torchvision==0.11.3+cu111 and torchaudio==0.10.2+cu11. You can modify the code of _ssim(...) function by inserting the following two lines at the beginning to solve this problem, but other problems will be encountered (in my case it is this one).

img1 = img1.unsqueeze(0)
img2 = img2.unsqueeze(0)

So the final solution should be using the correct versions. When I upgrade my cuda toolkit to 11.4 and recreate the conda env using conda env create --file ./environment.yml, everything is ok now. By the way, my system info is as following:

 ~  nvidia-smi
Tue Nov  5 10:26:21 2024       
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 470.182.03   Driver Version: 470.182.03   CUDA Version: 11.4     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|===============================+======================+======================|
|   0  NVIDIA GeForce ...  Off  | 00000000:65:00.0 Off |                  N/A |
| 30%   34C    P8    13W / 320W |   5648MiB / 10009MiB |      0%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+
                                                                               
+-----------------------------------------------------------------------------+
| Processes:                                                                  |
|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
|        ID   ID                                                   Usage      |
|=============================================================================|
|    0   N/A  N/A      7696      C   ...vs/scaffold_gs/bin/python     5461MiB |
|    0   N/A  N/A      7837      G   /usr/lib/xorg/Xorg                 18MiB |
|    0   N/A  N/A     10021      G   /usr/bin/gnome-shell               15MiB |
|    0   N/A  N/A     27031      G   /usr/lib/xorg/Xorg                112MiB |
|    0   N/A  N/A     27160      G   /usr/bin/gnome-shell               34MiB |
+-----------------------------------------------------------------------------+
 ~
 ~  nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2021 NVIDIA Corporation
Built on Mon_Oct_11_21:27:02_PDT_2021
Cuda compilation tools, release 11.4, V11.4.152
Build cuda_11.4.r11.4/compiler.30521435_0
 ~
 ~ conda list | grep torch
ffmpeg                    4.3                  hf484d3e_0    pytorch
pytorch                   1.12.1          py3.7_cuda11.6_cudnn8.3.2_0    pytorch
pytorch-mutex             1.0                        cuda    pytorch
pytorch-scatter           2.1.0           py37_torch_1.12.0_cu116    pyg
torchaudio                0.12.1               py37_cu116    pytorch
torchvision               0.13.1               py37_cu116    pytorch