PaliGemma 2
Google DeepMind — Gemma
partial detail 15 method fields published
Evaluation
Effect on task performance
relative improvement in transfer metrics when equipping PaliGemma 2 3B (224px2) with the bigger 9B LM while keeping resolution (3.7x more FLOPs), or keeping the model size and increasing resolution to 448px2 (4.6x more FLOPs); task groups sensitive to both, to model size, or to resolution
How it was run
Open a release to see the setup, scoring, and other details that the lab published. Only fields the lab actually disclosed are shown.