Jaswanth Bandlamudi

I still have this issue. I confirmed that the` os.path.join(DIR,'conf/data/object_detection/scannet.yaml') ` points to correct yaml.

I tried using s3dis_explorer.ipynb. ``` --------------------------------------------------------------------------- AttributeError Traceback (most recent call last) /home/jba/play_zone/torch-points3d/notebooks/s3dis_explorer.ipynb Cell 6' in () ----> 1[ dataset_options.data.dataroot = os.path.join(DIR,"data") ]()[2](vscode-notebook-cell://ssh-remote%2Briiico/home/jba/play_zone/torch-points3d/notebooks/s3dis_explorer.ipynb#ch0000005vscode-remote?line=1)[ dataset = S3DISFusedDataset(dataset_options.data) ]()[3](vscode-notebook-cell://ssh-remote%2Briiico/home/jba/play_zone/torch-points3d/notebooks/s3dis_explorer.ipynb#ch0000005vscode-remote?line=2)[ print(dataset) AttributeError: 'NoneType'...

I could solve the problem by using the following as input to Omegaconf.create() function ``` shapenet_yaml = """ task: panoptic class: s3dis.S3DISFusedDataset dataroot: %s fold: 5 first_subsampling: 0.04 grid_size: ${first_subsampling}...

The image size is defaulted in post_process_coords function to 1600*900 in case of nuscenes dataset, But Lyft dataset has two types of images one with 1224 * 1024 and 1920...

The website says 503 Unavailable. Is there any alternative link to download the dataset ??

I think the recent pymeshlab release (2022.2.post4) is broken in the case of e57 files. Please install as `pip install pymeshlab==2022.2.post2` It should work. It's weird that it affected working...

@alemuntoni I found the issue today. I haven't debugged it yet. But I knew pymeshlab supported e57 files. So I found the above solution.