Error when running training scripts about differentiable_volumetric_rendering HOT 14 CLOSED

I see, you have an error because of imageio.

I think adding
imageio.plugins.freeimage.download()
to e.g. train.py (after importing imageio) should fix the problem.

@kh9543, Thanks a lot for your help!

from differentiable_volumetric_rendering.

Ah, I have the same problem.
The error is from depth loading part with imageio.

I solve it by installing freeimage plugin of the imageio module.

from differentiable_volumetric_rendering.

Hi @MultiPath , thanks a lot for your interest!

Regarding your error, please check the detailed instructions where we indicate that the *_pretrained.yaml configs cannot be used for training. Can you try to start training with another config, e.g.

python train.py configs/single_view_reconstruction/multi_view_supervision/ours_combined.yaml

and check if this works for you? Thanks a lot.

from differentiable_volumetric_rendering.

@m-niemeyer Thanks for your quick reply!
All the configs showed the same errors on my side. It also reports IndexError: list index out of range even with ours_combined.yaml

Thanks

from differentiable_volumetric_rendering.

It first shows a lot errors of loading images, while I don't know why this happened:

Error occurred when loading field img of model b5ae5354fe7c166319cb07ecb5b4102 (03001627)
Error occurred when loading field img of model b9f28fa7c7181002d436d4e00775800d (02958343)
Error occurred when loading field img of model ca2294ffc664a55afab1bffbdecd7709 (03001627)
Error occurred when loading field img of model c393265168bcc56cf3ced7613d1c130b (04379243)
Error occurred when loading field img of model bb84310d8f022346f2b931201029bc76 (04379243)
Error occurred when loading field img of model bda24b02920f91d1887e8faf8f33b0c5 (03691459)
Error occurred when loading field img of model cf280956401024fe92225eae846f578f (03636649)
Error occurred when loading field img of model c91c78e14fccbab316a2b5862518c93 (04379243)
Error occurred when loading field img of model b23b29f4b08cb2e7b65b71636c46ae49 (04379243)
Error occurred when loading field img of model bd6a8b133fa4d269491d6cee03fef2a9 (03001627)
Error occurred when loading field img of model c627f3639caaf614a6589c01f678d8ea (04379243)
Error occurred when loading field img of model c87497b3c00b3116be8af56c801ecf41 (04256520)
Error occurred when loading field img of model ba673ea75085e46cbfd72d7396bc040a (03001627)
Error occurred when loading field img of model ba97b36204357858e9a890c7dd671d9a (02958343)
Error occurred when loading field img of model b5b6b09711cbee6daa44bfa127abe4bb (02958343)

from differentiable_volumetric_rendering.

The error messages you show just indicate that it cannot find the dataset. For the configs you tried, it needs the ShapeNet dataset in {your_github_clone}/data/ShapeNet. Can you check if the data is completely downloaded an extracted to this folder? Good luck!

from differentiable_volumetric_rendering.

Hi @m-niemeyer

I did run your scripts and it showed everything was downloaded and unzipped normally.
To give you a sense, the attached are some screenshots:

All the images were there.

from differentiable_volumetric_rendering.

I was running your scripts/download_dataset.sh

The same error also happened to me for NMR dataset. Both have these errors.

from differentiable_volumetric_rendering.

I pulled the repo but I can't reproduce the error.

From your screenshot I can see that the data is in ~/data/shapenet/differentiable_volumetric_rendering/data/ShapeNet. Did you clone the repo to ~/data/shapenet/differentiable_volumetric_rendering?

Please check again that the paths are correct. The data either has to be in {repo_clone}/data/ShapeNet or you can also adjust the dataset path directly in the config file..
Just to make sure, you can also pull the newest repo (maybe you pulled an earlier version). I will have a look if there is something wrong with the data if still not working.

from differentiable_volumetric_rendering.

@MultiPath Can you show the exception caught?
The error seems like coming from this line

from differentiable_volumetric_rendering.

Yes. I am sure the codes and paths are correct. I pulled the repo to ~/data/shapenet/differentiable_volumetric_rendering and run the training script under that folder.

Just to make sure if the uploaded datasets are complete or not?
I am not sure about that, for instance, the downloaded sizes of zip files are:

The downloading scripts will automatically unzip the datasets, right?

from differentiable_volumetric_rendering.

@MultiPath Can you show the exception caught?
The error seems like coming from this line

differentiable_volumetric_rendering/im2mesh/data/core.py

Line 164 in 8388d6c

except Exception:

Thanks @kh9543

I disabled the "self.no_except" .... and finally got the real Exception now:

when running for ours_combined.yaml

from differentiable_volumetric_rendering.

I see, you have an error because of imageio.

I think adding
imageio.plugins.freeimage.download()
to e.g. train.py (after importing imageio) should fix the problem.

@kh9543, Thanks a lot for your help!

It works! Thank you very much @m-niemeyer @kh9543 for your kind help!

from differentiable_volumetric_rendering.

I solve this problem by
conda install -c conda-forge freeimage

from differentiable_volumetric_rendering.