I was training a model with 1 train clip and 1 test clip but I accidentally had train_subset_ratio=0.8 from a previous run which I assume made n_test_clips=0. I believe this caused IndexError: index is out of bounds for axis 0 with size 0 because when I set back train_subset_ratio=0.5 the error went away. I'm not sure how 0 test clips is handled in the train script but it seems like it's not handled correctly?
I was training a model with 1 train clip and 1 test clip but I accidentally had
train_subset_ratio=0.8from a previous run which I assume maden_test_clips=0. I believe this causedIndexError: index is out of bounds for axis 0 with size 0because when I set backtrain_subset_ratio=0.5the error went away. I'm not sure how 0 test clips is handled in the train script but it seems like it's not handled correctly?