PaliGemma 2: A Family of Versatile VLMs for Transfer
Open paper
Benchmark
Model
Rank
Results
image-captioning-on-textcaps
PaliGemma 2 10B
#1
CIDEr: 157.7
visual-grounding-on-refcoco-val
PaliGemma 2 10B
#7
Accuracy: 78.2
visual-question-answering-vqa-on-textvqa-val
PaliGemma 2 10B
#18
Accuracy: 76.6
Rank counts only results with a code link.