arxiv:2610.18097

APT: Feedback-Driven Agentic Post-Training for Text-to-Image Generation

Published on Oct 13

· Submitted by

Sirui Zhang on Oct 15

Microsoft Research

Upvote

Authors:

Sirui Zhang ,

Fangyun Wei ,

Abstract

APT is an automated framework that adaptively post-trains text-to-image models by using agents to evaluate capabilities, diagnose failures, and generate targeted reinforcement learning data in a self-improving loop.

Generated by thinkingmachines/Inkling-Small

Post-training is crucial for improving text-to-image models, yet existing pipelines typically rely on manually curated training data and fixed training distributions. We introduce APT (Agentic Post-Training), a fully automated framework that adaptively post-trains a base model according to its evolving capability profile. APT coordinates three agents to construct evaluation sets from user requirements, diagnose model-specific strengths and failure modes, and generate reinforcement learning data with an adaptive prompt distribution that emphasizes underperforming capabilities. By repeatedly evaluating, diagnosing, and updating the generator, APT lets the current model's behavior shape its next training tasks, forming a feedback loop inspired by recursive self-improvement. For general-purpose settings without user-specified evaluation scenarios, we further develop an evaluation harness that enables broad and diverse evaluation set generation. Experiments across multiple base models, benchmarks, and generation architectures demonstrate that APT is model-agnostic, benchmark-agnostic, architecture-agnostic, and fully automated.

View arXiv page View PDF Project page GitHub 10 Add to collection

Community

zsr200901

Paper author Paper submitter about 22 hours ago

Better models. Smarter training data.
Different image models struggle with different things. A fixed training set cannot keep up with every model’s evolving strengths and weaknesses.
APT makes post-training adaptive. Three agents work together to evaluate the current model, diagnose its failure modes, and generate the training prompts it needs next. After each update, the loop starts again.APT (Agentic Post-Training) turns post-training into a fully automated, evaluation-driven adaptive loop. Given a user-specified evaluation scope, the Evaluation Set Generation Agent constructs a diverse evaluation set consisting of prompt–evaluation rubric pairs; the Diagnosis Agent evaluates the current model on these pairs to identify its strengths and failure modes; and the RL Data Generation Agent adaptively samples next-round training prompts, prioritizing underperforming capabilities. The updated model is then re-evaluated in the next round, allowing the training distribution to continuously adapt to the model's evolving capability profile. This design eliminates the need for manually curated post-training data and makes APT model-agnostic, benchmark-agnostic, architecture-agnostic, and fully automated.