Visual Question Answering on ScienceQA Image Set
92.53Top-1 AccLLaVa+GPT-4
Evaluation Results
| Method | Links | |
|---|---|---|
| LLaVa+GPT-4mode=specialist2023.08 | 92.53 | |
| LyricsBackbone=Vicuna-13B, Evaluation protocol=0-shot2023.12 | 71.1 | |
| IDEFICS-80BBackbone=LLaMA-65B, Evaluation protocol=0-shot2023.12 | 68.9 | |
| Qwen-VL-Chatzero-shot=true2023.08 | 68.2 | |
| Qwen-VLBackbone=Qwen-7B, Evaluation protocol=0-shot2023.12 | 68.2 | |
| Qwen-VLzero-shot=true, backbone=Qwen-7B2023.08 | 67.1 | |
| InstructBLIPzero-shot=true, backbone=Vicuna-13B2023.08 | 63.1 | |
| InstructBLIPBackbone=Vicuna-13B, Evaluation protocol=0-shot2023.12 | 63.1 | |
| BLIP-2zero-shot=true, backbone=Vicuna-13B2023.08 | 61 | |
| BLIP-2Backbone=Vicuna-13B, Evaluation protocol=0-shot2023.12 | 61 | |
| IDEFICS-9BBackbone=LLaMA-7B, Evaluation protocol=0-shot2023.12 | 44.2 |