Transformer
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Transformer Inference | 12-layer Transformer 1024 tokens (inference) | Speedup342.74 | 24 | |
| Jacobian computation | Transformer | Median Runtime (ms)0.28 | 8 | |
| Scaling Efficiency | Transformer 128 tokens | Scaling Efficiency (Linear Projection)93.29 | 5 | |
| Finding Optimal Elimination Order | Transformer | Number of Multiplications4,656 | 5 |
Showing 4 of 4 rows