Kernel Runtime Optimization on FlashInfer-Bench gqa_paged_decode (test)
11Correct CountAdaExplore
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| AdaExploreSearch Strategy=50-step MCTS, LLM=GPT-5-mini, Hardware=NVIDIA B2002026.04 | 11 | 34 | 0.682 | 18.17 | 0.15 |
| Method | Links | |||||
|---|---|---|---|---|---|---|
| AdaExploreSearch Strategy=50-step MCTS, LLM=GPT-5-mini, Hardware=NVIDIA B2002026.04 | 11 | 34 | 0.682 | 18.17 | 0.15 |