domains / families / data / tao-generate-image-grounding

tao-generate-image-grounding

separate

"Two-step image grounding pipeline: extracts referring expressions from (image, caption) pairs and grounds them

When to use

• Generate phrase-grounded training data for referring-expression and grounding models. The VLM acts as a "teacher" annotator: Step 0 extracts referring expressions from the caption while looking at the image; Step 1 returns one bbox set per…

Skill page

Source