Testing

[manza-ari]

Testing
Download our pre-trained model and move it to models.
Run python demo.py -demo /path/to/image [-loadModel /path/to/image].

Kindly please define again how to run testing side. I have downloaded the pre-trained model and moved to models folder. Then I have to run demo.py which is in src folder? Am I right?

[xingyizhou]

Yes, in src folder, run python demo.py -demo /path/to/image [-loadModel /path/to/image].

[xyz2357]

It seems that directly put model in src/ is a better idea, because of the code
model = torch.load('hgreg-3d.pth').cuda() in demo.py line 15 https://github.com/xingyizhou/pytorch-pose-hg-3d/blob/84ad44e7a8aa15307b9a371ce85b3dee8d5ad2dc/src/demo.py#L15

[manza-ari]

I dont know but demo.py is not working and giving error.

[xingyizhou]

It seems a pytorch installation issue. Please check the pytorch homepage.

[manza-ari]

Thank you so much for all your posts and help.

But still, I am unable solve the issue. Did you make this using Python 2.7?

[xingyizhou]

Yes, I am using python 2.7.

[udion]

I am getting the following error, seems like pre-trained model is corrupted?

Note that the hgreg-3d.pth is present in the same directory as demo.py

[xingyizhou]

Hi,
I have only tested my model in python 2. Python 3 is not supported yet. here is a python3 compactable version (not tested) which you can train the model yourself.

[manza-ari]

Hi udion,

I have resolved the issue by setting a virtual environment of Python 2 and it worked absolutely fine. It is an easy and convenient way to run the testing. Though the author has released for Python 3. Please ask me if I can help you with screenshots.

[xingyizhou]

Thanks @kanza-ali !

[bilawal-mahmood]

I am still getting this error as i change the code for python 3. can you help @kanza-ali

[bilawal-mahmood]

'ascii' codec can't decode byte 0xc3 in position 13: ordinal not in range(128)
I am still getting this error. can someone help me solve this. I am using windows and python 2 does not have pytorch. not even for virtual enviroment