LAION-400M: Open Dataset of CLIP-Filtered 400 Million Image-Text Pairs

Schuhmann, Christoph, Vencu, Richard, Beaumont, Romain · arXiv · 2021