Available now

Florence-2 Phrase Grounding: find phrases in an image

Florence-2 Phrase Grounding finds where the things named in your sentence appear in an image and draws a labeled box around each.

Model ID

Input

Fields marked * are required. Review the price before generating.

The image to be processed.

Uploading a file needs an account. Create account

JPEG, PNG, WebP · up to 25 MB

Text input for the task

Example. Edit it or write your own.

Price$0.60 / run

You sign up first. Your settings are kept.

About this model

Overview

Upload an image and type a caption, such as "a woman holding a red umbrella next to a bicycle". Microsoft's Florence-2 matches the phrases in it to parts of the picture and returns the image with boxes, plus each box's position and label.

What you can build

  • Checks that a generated image contains what its prompt asked for
  • Box coordinates for cropping named objects
  • Labeled training or review data from captions

Prompting tips

  • Write the caption as plain noun phrases; each phrase that names a visible thing becomes a box.
  • Use words that describe what is visible, such as color or position, to tell similar objects apart.

What to check

Compare each returned region with the phrase it represents. Check missed subjects and incorrectly matched locations.

Inputs

  • Image (required): The URL of the image to be processed.
  • Text Input (required): Text input for the task.

Output and limits

  • Small or dense text in the image can come out misspelled.
  • Hands, crowds and complex poses can show artifacts; generate a few variations and pick the best.
  • Some prompts or outputs can be refused by safety checks.

Price guide

  • up to: $0.60 / run