contrastive-active-learning
contrastive-active-learning copied to clipboard
problem when I run the "batch_bald" on sst-2 dataset
Hello, when I try to run the command python run_al.py --dataset_name sst-2 --acquisition batch_bald
, it throws an error that "IndexError: Dimension out of range (expected to be in range of [-1, 0], but got 1)" when evaluation the Dpool
. Specifically, it occurs when computing the entropy of sample.
The problem doesn't exist in other acquisition strategies.
Thank you.