Train, use and visualize an aesthetic score predictor ( how much people like on average an image ) based on a simple neural net that takes CLIP embeddings as inputs.
Link to the AVA training data ( already prepared) : https://drive.google.com/drive/folders/186XiniJup5Rt9FXsHiAGWhgWz-nmCK_r?usp=sharing
Visualizations of 100k images from LAION 400M with the model ava+logos-l14-linearMSE : http://captions.christoph-schuhmann.de/aesthetic_viz_laion_ava+logos_L14_100k-linearMSE.html