PaliGemma 1
Google DeepMind — Gemma
rich detail 17 method fields published
Evaluation
Causal masking and learning objective ablation
autoregressive masking placement (prefix-LM vs masking prefix/image tokens), application of next-token-prediction loss to suffix only, and task-prefix usage in Stage1 pretraining
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.