Video Understanding on Video-ChatGPT
2.663ScoreReDiPrune
Evaluation results
| Method | Links | |
|---|---|---|
| ReDiPruneBackbone=LLaVA-NeXT-Video-7B, Keep ratio=0.10, TFLOPs (ratio %)=2.712 (9.28)2026.03 | 2.663 | |
| DivPrune*Backbone=LLaVA-NeXT-Video-7B, Keep ratio=0.10, TFLOPs (ratio %)=2.712 (9.28)2026.03 | 2.637 | |
| CDPruner*Backbone=LLaVA-NeXT-Video-7B, Keep ratio=0.10, TFLOPs (ratio %)=2.712 (9.28)2026.03 | 2.629 | |
| Original*Backbone=LLaVA-NeXT-Video-7B, Keep ratio=0.10, TFLOPs (ratio %)=29.211 (100)2026.03 | 2.52 | |
| ReDiPruneBackbone=LLaVA 1.5-7B, Keep ratio=0.10, TFLOPs (ratio %)=1.174 (10.11)2026.03 | 2.24 | |
| DivPrune*Backbone=LLaVA 1.5-7B, Keep ratio=0.10, TFLOPs (ratio %)=1.174 (10.11)2026.03 | 2.18 | |
| Original*Backbone=LLaVA 1.5-7B, Keep ratio=0.10, TFLOPs (ratio %)=11.606 (100)2026.03 | 2.13 | |
| CDPruner*Backbone=LLaVA 1.5-7B, Keep ratio=0.10, TFLOPs (ratio %)=1.174 (10.11)2026.03 | 2.12 |