Table 6 Estimated testing time per batch.

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

Dataset

Without optimization

With optimization

Multihead (4 heads)

Multihead (8 heads)

Proposed model

SST-5

224

319

188

Twitter

129

155

114

Laptop14

135

288

161

Restaurant14

388

571

357

Restaurant15

409

575

368

Restaurant16

422

581

382

Back to article page