Inference Energy Consumption Estimation on Theoretical
0.1152FLOPs (Giga)SpikeATE
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| SpikeATEQuantization=Binary2026.01 | 0.1152 | 0.0059 | 1.8943 | |
| SpikeATEQuantization=Ternary2026.01 | 0.1152 | 0.0149 | 2.5946 | |
| DECNN2026.01 | 0.258 | — | 3.2256 | |
| SoftProtoE2026.01 | 0.258 | — | 3.2256 | |
| HAST2026.01 | 0.5232 | — | 6.5412 | |
| Seq2Seq4ATE2026.01 | 2.4888 | — | 31.1104 | |
| BERT-RC2026.01 | 7.6448 | — | 95.5599 | |
| BERT-PT2026.01 | 7.6451 | — | 95.5636 | |
| Self-Training2026.01 | 7.6451 | — | 95.5636 | |
| CDA2026.01 | 8.5409 | — | 106.7618 | |
| GPT-3.52026.01 | 314,000,000,000,000 | — | 3,920,000,000,000,000 |