Table 5 Estimated training time per epoch.

From: A deep sentiment model combining ALBERT-driven context and EHO-optimized architecture

Dataset

Multihead (4 heads)

Multihead (8 heads)

Proposed

SST-5

410

643

472

Twitter

366

457

388

Laptop14

417

600

473

Restaurant14

819

1082

886

Restaurant15

822

1116

893

Restaurant16

822

1120

907

Back to article page