search

uber-research / DeepPruner

DeepPruner: Learning Efficient Stereo Matching via Differentiable PatchMatch (ICCV 2019)
Other
354 stars 41 forks source link

finetune_kitti.py does not have "--save_dir" argument #5

Closed WZG3661 closed 4 years ago

WZG3661 commented 4 years ago

I want to give two pictures, one is the left and the other is right, to predict the depth, so how should I run your code? Should I run finetune_kitti.py? I don't understand your Readme.txt, is finetune_kitti.py to train and test a model? I have downloaded the pre-trained model and I just want to use it not to train it, should I run another .py docunment? Really hope you can help me.

ShivamDuggal4 commented 4 years ago

Hi @WZG3661 There was a typo in deeppruner readme under the evaluation command section. I have fixed it now. Just follow this command to run DeepPruner on any dataset. https://github.com/uber-research/DeepPruner/tree/master/deeppruner#evaluation-command-on-any-dataset

Let me know for any other issue.

WZG3661 commented 4 years ago

Hi @WZG3661 There was a typo in deeppruner readme under the evaluation command section. I have fixed it now. Just follow this command to run DeepPruner on any dataset. https://github.com/uber-research/DeepPruner/tree/master/deeppruner#evaluation-command-on-any-dataset

Let me know for any other issue.

Thanks a lot! But when I run the submission_kitti.py, I got another problem: `File "/home/lihua/Downloads/StereoMatching/DeepPruner-master/deeppruner/submission_kitti.py", line 106, in main imgL = processed(imgL).numpy()

File "/usr/local/lib/python3.6/dist-packages/torchvision/transforms/transforms.py", line 49, in call img = t(img)

File "/usr/local/lib/python3.6/dist-packages/torchvision/transforms/transforms.py", line 76, in call return F.to_tensor(pic)

File "/usr/local/lib/python3.6/dist-packages/torchvision/transforms/functional.py", line 48, in to_tensor img = torch.from_numpy(pic.transpose((2, 0, 1)))

ValueError: axes don't match array` My Pytorch version is 0.4.1, and my torchview version is 0.2.1, can you help me fix it?

WZG3661 commented 4 years ago

I find that if I give two gray images, the problem above will arise. So I give two color images, but another problem arised: tuple indices must be integers or slices, not tuple in the line 128 of submissiong_kitti.py.

ShivamDuggal4 commented 4 years ago

Hi @WZG3661 , it seems like your config file hasn't been set correctly for evaluation. https://github.com/uber-research/DeepPruner/blob/b13a254773843b92ee8bbfb5dbc18cf9a3078ea4/deeppruner/models/config.py#L31

Please change the "mode" to "evaluation" in order to successfully run submission_kitti.py

WZG3661 commented 4 years ago

Thaks a lot! Hope you can make it clear in Readme.txt.