Action Prediction on Mind2Web Cross-Domain
85.7Operation F1OmniParser
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| OmniParserHTML free=true, image=true, variant=LS + ID2024.08 | 85.7 | 45.5 | 42 | |
| SeeClickHTML free=true, image=true2024.08 | 84.8 | 23.2 | 20.8 | |
| Qwen-VLHTML free=true, image=true2024.08 | 84.3 | 14.1 | 12 | |
| OmniParserHTML free=true, image=true, variant=LS + GD2024.08 | 80.6 | 44.9 | 36.8 | |
| DUAL-VCRRanker=DUAL-VCR_VNEI-TXT+VIS, Action Predictor=DUAL-VCRPRED2024.02 | 73.3 | 33.2 | 32.5 | |
| DUAL-VCRRanker=DUAL-VCR_VNEI-TXT, Action Predictor=DUAL-VCRPRED2024.02 | 72.9 | 32.4 | 32 | |
| GPT-4V+textual choiceHTML free=false, image=true2024.08 | 69.3 | 42.4 | 36.8 | |
| MindActRanker=MINDACTRANK, Action Predictor=MINDACTPRED2024.02 | 66.6 | 31.5 | 31 | |
| MindActHTML free=false, image=false2024.08 | 66.5 | 42.1 | 39.6 | |
| GPT-3.5-TurboHTML free=false, image=false2024.08 | 52.8 | 21.6 | 18.6 | |
| GPT-4HTML free=false, image=false2024.08 | 46.5 | 37.1 | 26.4 | |
| MindAct (gen)HTML free=false, image=false2024.08 | 44.7 | 14.2 | 11.9 | |
| CogAgentHTML free=true, image=true2024.08 | 42 | 20.6 | 15.5 | |
| GPT-4V+somHTML free=false, image=true2024.08 | 30.1 | — | 20.3 |