Trained on a subset of the Flickr faces dataset. <ol><li>Entirely new, hand written captions.</li><li>Zero NSFW Images in Dataset</li><li>Only photographs (Flicker faces has images of dolls those where pruned)</li></ol>You can describe in natural language or Vit tagging.<ol><li>An ugly mature female with a square face.</li><li>mature, square face, 1girl, ugly</li></ol>This lora address some of the issues with zero shot image classification. Such as the 60-80% accuracy on race. The fact that it does not ascribe aesthetics or face characteristics. As well as other issues.

Files with SHA256: c3994b3889ffe1c96d47e8b864f4739c42347c6ec3017d2237e8f5cb0cbc0af4