Training Priors Predict Text-To-Image Model Performance

Open in new window