Multiple Context Learning Networks for Visual Question Answering

Table 4

The results of fine-tuning BERT on GQA validation set.

Model	lr×	All
N = 1, BERT	0.001	64.25
N = 1, BERT	0.01	64.63
N = 1, BERT	0.1	64.51
N = 2, BERT	0.01	65.22
N = 3, BERT	0.01	65.11
N = 2, LSTM	-	65.01