Hybrid Mamba-Transformer Inference on agentic_burst workload
46.91Goodput (req/s)avmp_dynamic_b128
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| avmp_dynamic_b128variant=avmp_dynamic_b1282026.05 | 46.91 | 1.83 | |
| fixed_dual_mr05variant=fixed_dual_mr052026.05 | 25.69 | — |
| Method | Links | ||
|---|---|---|---|
| avmp_dynamic_b128variant=avmp_dynamic_b1282026.05 | 46.91 | 1.83 | |
| fixed_dual_mr05variant=fixed_dual_mr052026.05 | 25.69 | — |