Model

Train Time

Inference Time

GPU Memory

BERT

1.00

1.00

<16

BERT + TextRank

1.96

1.96

16

BERT + Random

1.98

2.00

16

Longformer

12.05

11.92

32

ToBERT

1.19

1.70

32

CogLTX

104.52

12.53

<16

Mamba

6.49

8.35

64