MICROSOFT'S FLORENCE-2 STATE-OF-THE-ART VISION MODEL (HUGGING FACE HUB)

Summary from the TLDR AI newsletter

Microsoft has released an MIT-licensed set of small VLMs that dramatically outperform much larger models on captioning, bounding, and classification.