Action Sequence Generation on OmniACT (test)
34Action ScoreUGround-V1-7B
Evaluation results
| Method | Links | |
|---|---|---|
| UGround-V1-7BInputs=Image (SeeAct-V), Planner=GPT-4o2024.10 | 34 | |
| UGround-V1-2BInputs=Image (SeeAct-V), Planner=GPT-4o2024.10 | 32.9 | |
| UGroundInputs=Image (SeeAct-V), Planner=GPT-4o2024.10 | 32.8 | |
| UGroundInputs=Image (SeeAct-V), Planner=GPT-42024.10 | 31.1 | |
| SeeClickInputs=Image (SeeAct-V), Planner=GPT-4o2024.10 | 29.6 | |
| SeeClickInputs=Image (SeeAct-V), Planner=GPT-42024.10 | 28.9 | |
| DetACTInputs=Image + Text, Planner=GPT-42024.10 | 17 | |
| DetACTInputs=Text, Planner=GPT-42024.10 | 11.6 |