UGround is a universal visual grounding model for GUI agents that uses only screenshots to locate elements and outperforms existing agents despite lacking text-based inputs.
Due to the huge computation cost of handling high-resolution images, we use LoRA (Hu et al., 2022) for instruction finetuning in the two stages, with a device batch size of 4
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Navigating the Digital World as Humans Do: Universal Visual Grounding for GUI Agents
UGround is a universal visual grounding model for GUI agents that uses only screenshots to locate elements and outperforms existing agents despite lacking text-based inputs.