contrastive-u-dae / README.md
nightknocker's picture
Update README.md
82143b8 verified
|
raw
history blame
744 Bytes
metadata
license: apache-2.0
pipeline_tag: image-feature-extraction

Contrastive Booru-Image Embedding Model

The ViT-B model was replaced by another Encoder.

The images were tagged by the CL tagger. The caption (prompt) was tokenized by ModernBERT.

Objectives

Image-Caption Similarity Scoring: The model assigns a score to each image based on how closely it is related to its associated caption.

Image Labeling: The image embedding is used to label images by character name.

Updates

This version contains up to 4000 female characters extracted from public and private datasets.