Penguin-VL is a compact vision-language model family built to study how far multimodal efficiency can be pushed by redesigning the vision encoder, rather than only scaling data or model size.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results