Hugging Face's logo

Spaces:
Sijuade
/
CLIPSegmentation
Runtime error

App Files Community
CLIPSegmentation / README.md
Sijuade's picture
Sijuade
Update README.md
d193fbc
preview code
|
raw
history blame
1.46 kB
metadata
title: CLIPSegmentation
emoji: 🦀
colorFrom: red
colorTo: indigo
sdk: gradio
sdk_version: 3.46.0
app_file: app.py
pinned: false
license: mit

CLIP Segmentation

CLIP Segmentation Project leverages the power of OpenAI's CLIP model combined with a segmentation decoder to perform image segmentation based on textual prompts. Provide an image and a text prompt, and get segmented masks for each prompt.

Features

  • Textual Prompt Segmentation: Segment images based on textual prompts.
  • Multiple Prompts: Support for multiple prompts separated by commas.
  • Interactive UI: User-friendly interface for easy image uploads and prompt inputs.

Usage

  1. Upload an image using the provided interface.
  2. Enter your text prompts separated by commas.
  3. Click on "Visualize Segments" to get the segmented masks.
  4. Hover over a class to view the individual segment.

How It Works

The CLIP Segmentation Project combines the power of a pretrained CLIP model with a segmentation decoder. The CLIP model, developed by OpenAI, understands images paired with natural language. By combining this with a segmentation decoder, we can generate segmented masks for images based on textual prompts, bridging the gap between vision and language in a unique way.

Acknowledgements