| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| OpenAudioBench English subsets (test) | VocalNet-MDM | TPS2,153.43 | 15 | 3mo ago | |
| Long-context sequences 8K to 60K | HyLRA | Avg TTFT14.85 | 12 | 3mo ago | |
| Model Efficiency Benchmarking Llama3.1-8B | GradPruner | Training Time62.4 | 11 | 3mo ago | |
| Five Organisms | CDD | Mean Inference Time (s)13 | 10 | 8d ago | |
| Efficiency Profiling Workload G.7 Detailed Efficiency Results | LLMBoost | End-to-End Latency (s)13.75 | 9 | 3mo ago | |
| EgoHumans (test) | AHAP | Inference Time (s)0.34 | 7 | 2mo ago | |
| MME-RealWorld Lite | UniVLR | Average Time per Sample (s)3.2 | 5 | 21d ago | |
| HRBench8K | UniVLR | Average Time per Sample (s)4.5 | 5 | 21d ago | |
| NQ (test) | Inference Time (h)3.79 | 5 | 3mo ago | ||
| Efficiency Benchmark A40 4096 4096 | ChunkKV | Latency (s)160.32 | 4 | 23d ago | |
| 6 datasets (test) | CLIP (no defense) | Inference Time (s)3.63 | 4 | 1mo ago | |
| GPT-2 124M | Inference Speed (x)1 | 1 | 3mo ago |
