Issues creating endpoint files

[huannllls]

Hello author, I encountered a problem when using this program，TractSeg/blob/master/tractseg/libs/create_endpoints_mask_with_clustering.py，Most ".trk" are fine when generating endpoint files,However, some files report the following error when running:
Histogram of cluster labels:
[1176]
Labels of the two biggest clusters: [0]
Traceback (most recent call last):
File "/home/hls/Atext/TractSeg-master/TractSeg/tractseg/libs/create_endpoints_mask_with_clustering.py", line 116, in
points_start, points_end = cluster(points_subset, algorithm="DBSCAN")
File "/home/hls/Atext/TractSeg-master/TractSeg/tractseg/libs/create_endpoints_mask_with_clustering.py", line 59, in cluster
points_start, points_end = select_two_biggest_clusters(labels, points)
File "/home/hls/Atext/TractSeg-master/TractSeg/tractseg/libs/create_endpoints_mask_with_clustering.py", line 45, in select_two_biggest_clusters
cluster_B = points[labels == biggest_labels[1]]
IndexError: index 1 is out of bounds for axis 0 with size 1
This error resulted in me having 11 files with different numbers and only being able to generate 142 endpoint files。
907656 677968 856766 685058 792564 833249 715041 786569 814649 816653 622236
In addition, I learned that this project needs to train 3 parts of weights when training its own weights, but I am not sure what f1 value they need to achieve in the end. I don’t seem to have achieved a good result when training TOM weights and endpoint weights. The effect, theoretically what effect can each of them achieve?

Thank you very much for your answer！