Action Prediction on ActPred
0.56AccuracyTrimTokenator-LC
Evaluation results
| Method | Links | |
|---|---|---|
| TrimTokenator-LCBackbone=LLaVA-1.5-7B, Retention Ratio=0.22025.12 | 0.56 | |
| TrimTokenator-LCBase Model=InternVL-Chat-ViT-6B-Vicuna-13B2025.12 | 0.54 | |
| DivPruneBackbone=LLaVA-1.5-7B, Retention Ratio=0.22025.12 | 0.54 | |
| TrimTokenatorBackbone=LLaVA-1.5-7B, Retention Ratio=0.22025.12 | 0.54 | |
| SparseVLMBackbone=LLaVA-1.5-7B, Retention Ratio=0.22025.12 | 0.535 | |
| VisionZipBackbone=LLaVA-1.5-7B, Retention Ratio=0.22025.12 | 0.535 | |
| FastVBackbone=LLaVA-1.5-7B, Retention Ratio=0.22025.12 | 0.525 | |
| VisionZipBase Model=InternVL-Chat-ViT-6B-Vicuna-13B2025.12 | 0.52 | |
| DivPruneBase Model=InternVL-Chat-ViT-6B-Vicuna-13B2025.12 | 0.52 | |
| Original ModelBackbone=LLaVA-1.5-7B, Retention Ratio=1.02025.12 | 0.52 | |
| VTWBackbone=LLaVA-1.5-7B, Retention Ratio=0.22025.12 | 0.52 | |
| TrimTokenatorBase Model=InternVL-Chat-ViT-6B-Vicuna-13B2025.12 | 0.515 | |
| CATPBackbone=LLaVA-1.5-7B, Retention Ratio=0.22025.12 | 0.515 | |
| Original ModelBase Model=InternVL-Chat-ViT-6B-Vicuna-13B2025.12 | 0.51 | |
| CATPBase Model=InternVL-Chat-ViT-6B-Vicuna-13B2025.12 | 0.51 | |
| SparseVLMBase Model=InternVL-Chat-ViT-6B-Vicuna-13B2025.12 | 0.505 | |
| VTWBase Model=InternVL-Chat-ViT-6B-Vicuna-13B2025.12 | 0.49 | |
| TrimTokenator-LCBackbone=LLaVA-1.5-13B, Retention Ratio=0.22025.12 | 0.485 | |
| FastVBase Model=InternVL-Chat-ViT-6B-Vicuna-13B2025.12 | 0.48 | |
| DivPruneBackbone=LLaVA-1.5-13B, Retention Ratio=0.22025.12 | 0.475 | |
| SparseVLMBackbone=LLaVA-1.5-13B, Retention Ratio=0.22025.12 | 0.47 | |
| TrimTokenatorBackbone=LLaVA-1.5-13B, Retention Ratio=0.22025.12 | 0.47 | |
| VTWBackbone=LLaVA-1.5-13B, Retention Ratio=0.22025.12 | 0.465 | |
| VisionZipBackbone=LLaVA-1.5-13B, Retention Ratio=0.22025.12 | 0.465 | |
| FastVBackbone=LLaVA-1.5-13B, Retention Ratio=0.22025.12 | 0.46 | |
| Original ModelBackbone=LLaVA-1.5-13B, Retention Ratio=1.02025.12 | 0.45 | |
| CATPBackbone=LLaVA-1.5-13B, Retention Ratio=0.22025.12 | 0.445 | |
| Original ModelBackbone=Yi-VL-6B, Retention Ratio=1.02025.12 | 0.405 | |
| VisionZipBackbone=Yi-VL-6B, Retention Ratio=0.22025.12 | 0.405 | |
| TrimTokenator-LCBackbone=Yi-VL-6B, Retention Ratio=0.22025.12 | 0.405 | |
| VTWBackbone=Yi-VL-6B, Retention Ratio=0.22025.12 | 0.4 | |
| DivPruneBackbone=Yi-VL-6B, Retention Ratio=0.22025.12 | 0.4 | |
| SparseVLMBackbone=Yi-VL-6B, Retention Ratio=0.22025.12 | 0.395 | |
| CATPBackbone=Yi-VL-6B, Retention Ratio=0.22025.12 | 0.395 | |
| TrimTokenatorBackbone=Yi-VL-6B, Retention Ratio=0.22025.12 | 0.395 | |
| FastVBackbone=Yi-VL-6B, Retention Ratio=0.22025.12 | 0.39 |