Semantic Segmentation with ScanNet, help me.

[makotoniimi]

I want to Semantic Segmentation with ScanNet.
I used the article as a reference and the Semantic Segmantation using SemanticKITTI was successful.
Then I modified that program and tried to do Semantic Segmantation with ScanNet and failed.
I got the following error in the attached program.
semantic_torch-scannet.py.txt

Traceback (most recent call last):
File "semantic_torch-scannet.py", line 141, in
data = test_split.get_data(pc_idx)
File "/home/bignum/Open3D-ML/.venv/lib/python3.8/site-packages/open3d/_ml3d/datasets/scannet.py", line 177, in get_data
scene = self.path_list[idx]
IndexError: list index out of range

It seems to be unable to read the ScanNet dataset.
What am I doing wrong?

I have downloaded over 1TB of ScanNet datasets and have unzipped the zip files in each directory.
The file tree for my ScanNet Dataset is attached.
tree.txt

[makotoniimi]

I noticed that my ScanNet dataset does not have .npy files.
How do I create these .npy files?

[makotoniimi]

I ran script/preprocess_scannet.py and got a .npy files.
I changed the dataset_path to ScanNetData/scans.
But still get the same error.
Can someone please help me?