AI Paper Atlas

Research map / Reinforcement learning / RL from feedback

Policy-Based Tuning of Autoregressive Image Models with Instance- and Distribution-Level Rewards

Orhun Bugra Baran, Melih Kandemir, Ramazan Gokberk Cinbis

ECCV 2026 · Reinforcement learning › RL from feedback · approach: RL fine-tuning of generative models · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
fine-tune with CLIP, HPSv2 and LOO-FID rewards via GRPOLlamaGen-L · ImageNet 256 (CFG 1.5)FID: 4.64 -> 3.83Improved
fine-tune with proposed rewards without CFGLlamaGen-XL · ImageNet 256 (no CFG)FID: 12.66 -> 4.55Improved
fine-tune with proposed rewards without CFGLlamaGen-B · ImageNet 256 (no CFG)FID: 20.89 -> 8.91Improved
fine-tune with proposed rewardsLlamaGen-B · ImageNet 256 (CFG 1.5)Recall: 0.61 -> 0.54Hurt
apply method to VQGAN backbone (no top-k)VQGAN · ImageNet 256FID: 21.90 -> 15.47Improved
add LOO-FID and adaptive entropy versus AR-GRPO (CLIP+HPSv2)LlamaGen-L · ImageNet 256 (CFG 1.5)FID: 6.37 -> 3.83Improved
add LOO-FID reward to CLIP+HPS rewards (ablation)LlamaGen-L · ImageNet 256 (CFG 1.5)FID: 6.21 -> 3.83Improved
use CLIP and HPS rewards together versus HPS only (ablation)LlamaGen-L (no FID reward) · ImageNet 256IS: 176.73 -> 220.28Improved
add KL regularization weight 3.0 versus none (ablation)LlamaGen-L · ImageNet 256FID: 10.14 -> 3.83Improved
add KL regularization to FID-only reward (ablation)LlamaGen-L FID-only · ImageNet 256FID: 3.97 -> 4.07No clear effect
add entropy regularization 4e-3 versus none (ablation)LlamaGen-L · ImageNet 256FID: 4.96 -> 3.83Improved
use EMA statistics instead of batch statistics for LOO-FID (ablation)LlamaGen-L · ImageNet 256FID: 4.06 -> 3.83Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy