Attention Latency on Qwen3-30B-A3B-Instruct-2507 (64K Context)
1.07Speedup Factor
CompactAttention-FP
Updated 2mo ago
Evaluation Results
| Method | Links | |
|---|---|---|
2026.05 | 1.07 | |
2026.05 | 1 | |
2026.05 | 0.91 | |
2026.05 | 0.85 |
CompactAttention-FP
| Method | Links | |
|---|---|---|
2026.05 | 1.07 | |
2026.05 | 1 | |
2026.05 | 0.91 | |
2026.05 | 0.85 |