CLIP with Quality Captions: A Strong Pretraining for Vision Tasks