Couldn't run the code

[yuchenlichuck]

Here is what I run after train_network.py +dataset=[cars] I try to run only shapenet dataset, but failed, could you give me some advise? Thank you

/home/dubaiprince/miniconda3/envs/gaussian_splatting/bin/python /home/dubaiprince/Projects/splatter-image/train_network.py +dataset=[cars] 
/home/dubaiprince/miniconda3/envs/gaussian_splatting/lib/python3.7/site-packages/torchvision/io/image.py:13: UserWarning: Failed to load image Python extension: /home/dubaiprince/miniconda3/envs/gaussian_splatting/lib/python3.7/site-packages/torchvision/image.so: undefined symbol: _ZN2at4_ops19empty_memory_format4callEN3c108ArrayRefIlEENS2_8optionalINS2_10ScalarTypeEEENS5_INS2_6LayoutEEENS5_INS2_6DeviceEEENS5_IbEENS5_INS2_12MemoryFormatEEE
  warn(f"Failed to load image Python extension: {e}")
wandb: Currently logged in as: liyc. Use `wandb login --relogin` to force relogin
wandb: wandb version 0.16.1 is available!  To upgrade, please run:
wandb:  $ pip install wandb --upgrade
wandb: Tracking run with wandb version 0.15.10
wandb: Run data is saved locally in /home/dubaiprince/Projects/splatter-image/experiments_out/2024-01-09/12-30-03/wandb/run-20240109_123004-qnzcmbqp
wandb: Run `wandb offline` to turn off syncing.
wandb: Syncing run earthy-thunder-11
wandb: ⭐️ View project at https://wandb.ai/liyc/gs_pred
wandb: 🚀 View run at https://wandb.ai/liyc/gs_pred/runs/qnzcmbqp
Setting up [LPIPS] perceptual loss: trunk [vgg], v[0.1], spatial [off] [09/01 12:30:09]
/home/dubaiprince/miniconda3/envs/gaussian_splatting/lib/python3.7/site-packages/torchvision/models/_utils.py:209: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and will be removed in 0.15, please use 'weights' instead.
  f"The parameter '{pretrained_param}' is deprecated since 0.13 and will be removed in 0.15, "
/home/dubaiprince/miniconda3/envs/gaussian_splatting/lib/python3.7/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and will be removed in 0.15. The current behavior is equivalent to passing `weights=VGG16_Weights.IMAGENET1K_V1`. You can also use `weights=VGG16_Weights.DEFAULT` to get the most up-to-date weights.
  warnings.warn(msg)
Loading model from: /home/dubaiprince/miniconda3/envs/gaussian_splatting/lib/python3.7/site-packages/lpips/weights/v0.1/vgg.pth [09/01 12:30:10]
Error executing job with overrides: ['+dataset=[cars]']
Traceback (most recent call last):
  File "/home/dubaiprince/Projects/splatter-image/train_network.py", line 131, in main
    shuffle=True)
  File "/home/dubaiprince/miniconda3/envs/gaussian_splatting/lib/python3.7/site-packages/torch/utils/data/dataloader.py", line 344, in __init__
    sampler = RandomSampler(dataset, generator=generator)  # type: ignore[arg-type]
  File "/home/dubaiprince/miniconda3/envs/gaussian_splatting/lib/python3.7/site-packages/torch/utils/data/sampler.py", line 108, in __init__
    "value, but got num_samples={}".format(self.num_samples))
ValueError: num_samples should be a positive integer value, but got num_samples=0

[mengxuyiGit]

same problem here. Seems that the data format expected by dataloader is different from that of original SRN dataset. This happens to CO3D dataset too. Could the authors provide a processed dataset?

[13253318778]

same problem here. I relly don't know how to solve the problem,so I would like to seek help from everyone and the author. Thank you very much in advance!

[Liangym1225]

The file path is wrong. You need to modify line22 in /scene/srn.py to self.base_path = os.path.join(SHAPENET_DATASET_ROOT, "srn_{}/{}_{}".format(cfg.data.category, cfg.data.category, dataset_name))

[13253318778]

The file path is wrong. You need to modify line22 in /scene/srn.py to self.base_path = os.path.join(SHAPENET_DATASET_ROOT, "srn_{}/{}_{}".format(cfg.data.category, cfg.data.category, dataset_name))

Thank you very much for your help! I made the changes as you suggested, but the error still persists.

[Liangym1225]

@13253318778 Maybe your SHAPENET_DATASET_ROOT is not correct. srn_{cars,chairs} should be placed in SHAPENET_DATASET_ROOT.

[13253318778]

@13253318778 Maybe your SHAPENET_DATASET_ROOT is not correct. srn_{cars,chairs} should be placed in SHAPENET_DATASET_ROOT.

I sincerely thanks for your help, but I need to bother you again. Can you please take a look at my dataset path? Is this what you mean? My dataset path in the code is like this: SHAPENET_DATASET_ROOT = r"G:\dataset\NMR_Dataset"

[Liangym1225]

The path looks good but you should download src_*.zip. #9 will be helpful.

[13253318778]

路径看起来不错，但您应该下载 .#9 会有所帮助。src_*.zip

I am truly, deeply grateful for your help. The best wishes to you.

[yuchenlichuck]

yes I fix it the srn_{cars,chairs} should be placed in SHAPENET_DATASET_ROOT.

[yangeer]

The path looks good but you should download src_*.zip. #9 will be helpful.

Thank you very much for your help, can you also help me to take a look at my dataset path?

[yuchenlichuck]

change it you will get it

[szymanowiczs]

@yangeer maybe you can print what self.intrins finds? Make sure to download the srn_*.zip file. I also updated readme - hopefully it'll help.

[Li-colonel]

I think the issue is caused by a repeated line here

[yangeer]

Thank you very much for your help, I Ifinally solved the problem，I realized that the files in my google cloud drive were empty

[yuedajiong]

say thanks to me, I can fix this bug: :-)

1. SHAPENET_DATASET_ROOT='/root/superi/splatter/data/SRN/'

   absulate path, not relative path, for glob. PLEASE!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

2. self.base_path = os.path.join(SHAPENET_DATASETROOT, "srn{}/{}_{}".format(cfg.data.category, cfg.data.category,

   cfg.data.category,

                                                                         dataset_name))

   comment one-line: #cfg.data.category,

of course, you can: print('hi, here:', self.base_path, self.intrins)

have fun.