From: Survey of transformers and towards ensemble learning using transformers for natural language processing
Model
w1
w2
w3
w4
Weight
0.938
0.951
0.894
0.964