Hybrid Mamba-Transformer Inference on mixed_long workload
65.07Goodput (req/s)avmp_dynamic_b128
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| avmp_dynamic_b128variant=avmp_dynamic_b1282026.05 | 65.07 | 2.39 | |
| fixed_dual_mr05variant=fixed_dual_mr052026.05 | 27.25 | — |
| Method | Links | ||
|---|---|---|---|
| avmp_dynamic_b128variant=avmp_dynamic_b1282026.05 | 65.07 | 2.39 | |
| fixed_dual_mr05variant=fixed_dual_mr052026.05 | 27.25 | — |