CommerceGen-Studio: An End-to-End Framework for Commercial-Grade Main-Image Generation
CommerceGen-Studio: An End-to-End Framework for Commercial-Grade Main-Image Generation
CommerceGen-Studio 以三階段流程生成商用級電商商品主圖,結合迭代式中繼資料蒸餾、版面最佳化及專用擴散模型,以維持主體保真度、商標完整性和文字準確度。框架每天生成數百萬張主圖,每張延遲數秒,並推出 CommerceGen-Bench;在兩種編輯測試中,其整體分數均高於受測開源基線,且在指令遵循和產品融合方面勝過數個閉源模型。
Published on Oct 14
Authors:
,
,
,
,
,
,
,
,
,
Abstract
CommerceGen-Studio is a three-stage framework that generates commercial e-commerce product images by combining iterative metadata distillation, layout optimization, and a specialized diffusion model to maintain subject fidelity, trademark integrity, and text accuracy at production scale.
We present CommerceGen-Studio, an end-to-end framework for generating commercial-grade main images for e-commerce Stock Keeping Units (SKUs). Our framework is organized along a three-stage pipeline of information extraction, semantic rewriting, and multi-condition generation: Prompt-Eval-Refine-Loop Agent (PERL-Agent) distills heterogeneous SKU metadata into concise selling-point text lists through an iterative closed loop of VLM and LLM roles scored by a reference-free rubric. Layout Thinker Optimization (LTO) aligns the rewriter with both caption and rendered-image quality via alternating domain-specific policy optimization. SKU-DiT, trained under Reinforce-Distill-Reinforce (RDR) progressive training, conditions jointly on the white-background product image, the trademark crop, and the structured layout to hold subject fidelity, trademark integrity, and character-level text accuracy together. The framework runs in production at a scale of millions of main images per day with a latency of a few seconds per image, which we believe makes it among the first systems of this kind to operate at commercial scale. We also release CommerceGen-Bench, a benchmark of hard cases collected from live traffic across all product categories. On it, our framework achieves the highest overall scores among the evaluated open-source baselines on both editing tracks, with and without prompt enhancement. It also matches much larger closed-source models, and beats several of them at Instruction Following and Product Integration - the two properties that determine whether an image can be published.
View arXiv page View PDF Add to collection
Get this paper in your agent:
hf papers read 2610.20603
Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash
Models citing this paper 0
No model linking this paper
Cite arxiv.org/abs/2610.20603 in a model README.md to link it from this page.
Datasets citing this paper 0
No dataset linking this paper
Cite arxiv.org/abs/2610.20603 in a dataset README.md to link it from this page.
Spaces citing this paper 0
No Space linking this paper
Cite arxiv.org/abs/2610.20603 in a Space README.md to link it from this page.
Collections including this paper 0
No Collection including this paper
Add this paper to a collection to link it from this page.
來源:HuggingFace Daily Papers(社區熱門論文) · huggingface.co