Issues with the Stanford demo

[amineremache]

Hello,
I tried to run the Stanford demo, and I got multiple issues:
1- When you run the python -m demo.stanford command for the first time (you don't have the .ply file yet) it should run the download function since the file_name doesn't match with any local file.

But then, you get this error:

Because the defined file_name in the start of the demo file doesn't match the name of the file on the server-side if you follow the static part of the (conferenceRoom_1.ply vs 1.ply).

2- When you change the name on the code to 1.ply, it downloads the file, but from what I understood, it is not the good file.
Because then I got this error:

Which is caused by this line:

And when you see the structure of the data here:

It really has no "label" property, and I think the file is from ScanNet because when you visualize it you get this:

Sorry if this is not the proper way to write an issue, it is my first one, and I tried to be the most explicit I could.

[ghost]

Hi,

That's because 1.ply is from Scannet dataset & doesn't have the label attribute. What you need is a test file from S3DIS dataset.

Another hack would be to just pass torch.ones(SHAPE) as labels variable in stanford.py, it will run.

[amineremache]

Hi,

That's because 1.ply is from Scannet dataset & doesn't have the label attribute. What you need is a test file from S3DIS dataset.

Another hack would be to just pass torch.ones(SHAPE) as labels variable in stanford.py, it will run.

Yeah, I was ( still ? xD ) a "noob" back then, it was my first few days with this code base, but things are good now, thanks for your answer.