In autoregressive image generation, DPO is stronger on in-domain prompts, GRPO generalizes better out-of-domain, and reward model generalization partly shapes both.
MaskGIT: Masked generative image transformer
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Delving into RL for Image Generation with CoT: A Study on DPO vs. GRPO
In autoregressive image generation, DPO is stronger on in-domain prompts, GRPO generalizes better out-of-domain, and reward model generalization partly shapes both.