Inference Latency on BERT-base
40.54Attention Layer Latency (s)
EncFormer
Updated 4d ago
Evaluation Results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
2026.04 | 40.54 | 31.1 | 1.88 | 38.16 | 0.98 | 32.5 | 3.38 | 30.27 | 0.92 | 0 | 179.73 | |
2026.04 | 40.54 | 31.1 | 2.89 | 38.16 | 1.23 | 32.5 | 6.24 | 30.27 | 1.22 | 0 | 184.15 | |
2026.04 | 40.54 | 31.1 | 4.43 | 38.16 | 1.5 | 32.5 | 14.64 | 30.27 | 1.57 | 0 | 194.71 | |
2026.04 | 40.54 | 31.1 | 8.52 | 38.16 | 3.45 | 32.5 | 29.98 | 30.27 | 3.7 | 0 | 218.22 | |
2026.04 | 49.77 | 32.53 | 15.53 | 48.06 | 7.13 | 49.8 | 29.42 | 49.19 | 4.1 | 337.86 | 623.39 | |
2026.04 | 57.65 | 28.76 | 0.75 | 41.49 | 0.37 | 59.21 | 8.31 | 43.77 | 0.3 | 103.72 | 344.33 |