Video-Language Understanding on ActivityNet
46.36AccuracyOriginal
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| OriginalBit=16, Backbone=LLaVA-Next-Video-7B2025.03 | 46.36 | 2.58 | |
| CLSE-MEvaluator=Gemini-2.5-Flash, Pruning Strategy=w/ Token Merging, Retained Tokens=1942026.06 | 45 | 2.24 | |
| FiCoCo-VEvaluator=Gemini-2.5-Flash, Pruning Strategy=Hard Pruning, Retained Tokens=1942026.06 | 44.9 | 2.2 | |
| CLSE-MEvaluator=GPT-4o-mini, Pruning Strategy=w/ Token Merging, Retained Tokens=1942026.06 | 44.6 | 2.25 | |
| SparseVLMEvaluator=Gemini-2.5-Flash, Pruning Strategy=w/ Token Merging, Retained Tokens=1942026.06 | 44 | 2.2 | |
| SparseVLMEvaluator=GPT-4o-mini, Pruning Strategy=w/ Token Merging, Retained Tokens=1942026.06 | 43.6 | 2.21 | |
| OccamToken2026.05 | 43.5 | — | |
| CASPQUIP#Bit=2, Backbone=LLaVA-Next-Video-7B2025.03 | 43.12 | 2.4 | |
| Video-LLaVA2026.05 | 43.1 | — | |
| FiCoCo-VEvaluator=GPT-4o-mini, Pruning Strategy=Hard Pruning, Retained Tokens=1942026.06 | 43.1 | 2.23 | |
| VisionZip2026.05 | 43 | — | |
| SparseVLM2026.05 | 42.6 | — | |
| Video-LLaVA-7BEvaluator=Gemini-2.5-Flash, Pruning Strategy=None, Retained Tokens=20482026.06 | 42.5 | 2.12 | |
| Video-LLaVA-7BEvaluator=GPT-4o-mini, Pruning Strategy=None, Retained Tokens=20482026.06 | 42 | 2.2 | |
| QuIP#Bit=2, Backbone=LLaVA-Next-Video-7B2025.03 | 40.39 | 2.23 | |
| CLSEEvaluator=GPT-4o-mini, Pruning Strategy=Hard Pruning, Retained Tokens=1942026.06 | 39.9 | 2.03 | |
| CLSEEvaluator=Gemini-2.5-Flash, Pruning Strategy=Hard Pruning, Retained Tokens=1942026.06 | 39.9 | 1.99 | |
| PDropEvaluator=Gemini-2.5-Flash, Pruning Strategy=Hard Pruning, Retained Tokens=1942026.06 | 39.4 | 1.93 | |
| CASPGPTQBit=2.2, Backbone=LLaVA-Next-Video-7B2025.03 | 38.3 | 2.2 | |
| PDropEvaluator=GPT-4o-mini, Pruning Strategy=Hard Pruning, Retained Tokens=1942026.06 | 34.3 | 1.78 | |
| FastVEvaluator=GPT-4o-mini, Pruning Strategy=Hard Pruning, Retained Tokens=1942026.06 | 34.2 | 1.83 | |
| FastVEvaluator=Gemini-2.5-Flash, Pruning Strategy=Hard Pruning, Retained Tokens=1942026.06 | 34.2 | 1.74 | |
| CASPAQLMBit=2, Backbone=LLaVA-Next-Video-7B2025.03 | 31.3 | 1.73 | |
| CapRL-Video-178KBackbone=Molmo2-8B2026.06 | 29.8 | — | |
| CapRL-Video-178KBackbone=Molmo2-4B2026.06 | 23.1 | — | |
| AQLMBit=2, Backbone=LLaVA-Next-Video-7B2025.03 | 22.54 | 1.13 | |
| Tarsier2-7B-Video-178KBackbone=Molmo2-8B2026.06 | 18.7 | — | |
| ShareGPT4Video-8B-Video-178KBackbone=Molmo2-8B2026.06 | 17.9 | — | |
| GPT-4o-Video-178KBackbone=Molmo2-8B2026.06 | 17.3 | — | |
| GPTQBit=2.2, Backbone=LLaVA-Next-Video-7B2025.03 | 15.02 | 1 | |
| Tarsier2-7B-Video-178KBackbone=Molmo2-4B2026.06 | 11.2 | — | |
| ShareGPT4Video-8B-Video-178KBackbone=Molmo2-4B2026.06 | 10.7 | — | |
| GPT-4o-Video-178KBackbone=Molmo2-4B2026.06 | 10.2 | — |