提出树状混合剪枝框架，让大模型变小还能跑在单张显卡上。
TMP: Tree-structured Mixed-policy Pruning for Large-scale Image Generation and Editing
- 用树状结构整合多种剪枝策略，适配图文生成与扩散模型。
- 将800亿参数模型压缩至200亿，减少75%参数量，质量损失小。
- 可部署于单张24GB显卡，适合资源有限的研究者与开发者。
当前图像生成模型规模持续扩大以实现高保真合成，但其庞大的参数量与计算开销导致资源需求激增、显存占用过高。本文提出首个树状混合策略剪枝框架TMP，通用性强，适用于主流图像任务（T2I与TI2I）及架构（MoE与DiT），可应用于步骤蒸馏模型并作为最后优化阶段。我们在开源的SOTA模型HunyuanImage-3.0 instruct与高效模型Z-Image turbo上进行实验。所提方法成功将HunyuanImage-3.0从800亿参数压缩至200亿（75%压缩比），生成质量下降有限；并通过工程优化，使200亿参数版本可在单张24GB 4090 GPU上推理。相关代码与权重已集成至HunyuanImage-3.0的GitHub与HuggingFace仓库。此外，我们验证了该框架对Z-Image turbo的有效性：从60亿压缩至40亿（33%压缩比），性能几乎无损。
原文摘要 · Abstract (English)
Modern image generation model rapidly grows their sizes to meet high-fidelity image synthesis. However, they gradually become unaffordable for their enormous parameter consumption and computation budget that lead to massive resources requirement and gpu memory footprint. In this paper, we propose TMP, the first Tree-structured Mixed-policy Pruning framework that generalizes prevalent image tasks (T2I and TI2I) and architectures (Mixture-of-Experts (MoE) and Diffusion transformer (DiT)). It could be applied to the step-distilled models and contribute as the last stage. We perform experiments upon current open-sourced SOTA HunyuanImage-3.0 instruct and a popular efficient model Z-Image turbo. The proposed pruning framework manages to compress HunyuanImage 3.0 from 80B to 20B parameters at 75% reduction ratio, sacrificing limited generation quality. We also optimize to enable the inference of the pruned 20B version of HunyuanImage 3.0 on a single 24GB 4090 GPU by engineering skills. The inference script and model weight have been integrated into the existing HunyuanImage3.0 open-source github and huggingface repository. Besides, we prove the efficacy of TMP by compressing Z-Image turbo from 6B to 4B (33% reduction) with negligible degradation.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。