Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fix Dib-Renderer ImageRecon example #293

Open
wants to merge 2 commits into
base: master
Choose a base branch
from

Conversation

Shubhamjain27
Copy link

@Shubhamjain27 Shubhamjain27 commented Jul 11, 2020

Fixed the training, eval, and collate_function code to match updates to the ShapeNet class.

Related Issues:
#289
#170
#146
#25

Tests:
Ubuntu 18.04
CUDA 10.2
Pytorch 1.3

  • Ran python script to completion (including visualization).

Signed-off-by: Shubhamjain27 <[email protected]>
Signed-off-by: Shubhamjain27 <[email protected]>
@archana1998
Copy link

Hi Shubham, just a small doubt. If you've run the code to complete visualization, then can you please specify the number of iterations you used for the training to get a good result?

@zyz-notebooks
Copy link

Fixed the training, eval, and collate_function code to match updates to the ShapeNet class.

Related Issues:
#289
#170
#146
#25

Tests:
Ubuntu 18.04
CUDA 10.2
Pytorch 1.3

  • Ran python script to completion (including visualization).

After I modified it according to your file, the following problems appeared in the running evaluation. Have you encountered it before? Is there a good solution?
(kaolin_test) zyz@ZYZ:~/Project/kaolin/examples/ImageRecon/Dib$ python eval.py
converting to voxels: 100%|██████████████| 2034/2034 [00:00<00:00, 29123.63it/s]
converting to surface meshes: 100%|██████| 2034/2034 [00:00<00:00, 29205.99it/s]
converting to points: 100%|██████████████| 2034/2034 [00:00<00:00, 93667.26it/s]
Traceback (most recent call last):
File "eval.py", line 54, in
cache_dir='cache/',train = False, split = .7)
TypeError: init() got an unexpected keyword argument 'cache_dir'

After I remove cache_dir, it's following problem
(kaolin_test) zyz@ZYZ:~/Project/kaolin/examples/ImageRecon/Dib$ python eval.py
converting to voxels: 100%|██████████████| 2034/2034 [00:00<00:00, 28294.12it/s]
converting to surface meshes: 100%|██████| 2034/2034 [00:00<00:00, 27876.14it/s]
converting to points: 100%|██████████████| 2034/2034 [00:00<00:00, 92047.24it/s]
Traceback (most recent call last):
File "eval.py", line 56, in
valid_set = kal.dataloader.ShapeNet.Combination([points_set_valid, images_set_valid, meshes_set_valid], root='/media/zyz/Project/DataSET/ShapeNetCore_v1/')
AttributeError: module 'kaolin' has no attribute 'dataloader'

@Shubhamjain27
Copy link
Author

Hi @zyz-1998 , yes this dataloder will not work. The api has been update to kal.datasets. I have incorporated these changes in my pull request. Hope that helps.

@zyz-notebooks
Copy link

Hi @zyz-1998 , yes this dataloder will not work. The api has been update to kal.datasets. I have incorporated these changes in my pull request. Hope that helps.

Thank you very much for your help. I conducted the evaluation smoothly, but it seems that I did not get useful information in the terminal, such as the inability to visualize

@Shubhamjain27
Copy link
Author

In the eval code, there is visualisation code as well. You have to set the args.vis to True. Also, the images and meshes can be saved in case running on EC2 machine.

@cduguet
Copy link

cduguet commented Feb 8, 2021

Thanks, @Shubhamjain27! What version of ShapeNet did you use to train DIB-R? If I use ShapeNetCore.v2 I'm missing the models (I only find normalized models) and the renders. If you had to put it up together, how did you do it?
Thank you!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants