Multi-modal Understanding on TVL Benchmark
6.16SSVTP ScoreTVL-LLaMA (ViT-Base)
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| TVL-LLaMA (ViT-Base)Vision Modality=true, Tactile Modality=true, Language Modality=true2026.04 | 6.16 | 4.89 | 5.03 | |
| TVL-LLaMA (ViT-Tiny)Vision Modality=true, Tactile Modality=true, Language Modality=true2026.04 | 6.09 | 4.79 | 4.94 | |
| VitaTouchVision Modality=true, Tactile Modality=true, Language Modality=true2026.04 | 5.83 | 6.1 | 6.09 | |
| TVL-LLaMA (ViT-Small)Vision Modality=true, Tactile Modality=true, Language Modality=true2026.04 | 5.81 | 4.77 | 4.89 | |
| GPT-4VVision Modality=true, Tactile Modality=false, Language Modality=true2026.04 | 5.02 | 4.42 | 4.49 | |
| ViP-LLaVA 13BVision Modality=true, Tactile Modality=false, Language Modality=true2026.04 | 4.1 | 3.76 | 3.8 | |
| LLaVA-1.5 7BVision Modality=true, Tactile Modality=false, Language Modality=true2026.04 | 3.64 | 3.55 | 3.56 | |
| LLaVA-1.5 13BVision Modality=true, Tactile Modality=false, Language Modality=true2026.04 | 3.55 | 3.63 | 3.62 | |
| ViP-LLaVA 7BVision Modality=true, Tactile Modality=false, Language Modality=true2026.04 | 2.72 | 3.44 | 3.36 | |
| SSVTP-LLaMAVision Modality=true, Tactile Modality=true, Language Modality=false2026.04 | 2.58 | 3.67 | 3.54 | |
| LLaMA-AdapterVision Modality=true, Tactile Modality=false, Language Modality=true2026.04 | 2.56 | 3.08 | 3.02 | |
| BLIP-2 Opt-6.7bVision Modality=true, Tactile Modality=false, Language Modality=true2026.04 | 2.02 | 2.72 | 2.64 | |
| InstructBLIP 13BVision Modality=true, Tactile Modality=false, Language Modality=true2026.04 | 1.44 | 1.21 | 1.24 | |
| InstructBLIP 7BVision Modality=true, Tactile Modality=false, Language Modality=true2026.04 | 1.4 | 1.3 | 1.31 |