<![CDATA[沉迷单车的追风少年]]>

<![CDATA[沉迷单车的追风少年]]>https://blog.youkuaiyun.com/qq_41895747zh-cnhttps://blog.youkuaiyun.com/<![CDATA[Seedance2.0 技术报告：Advancing Video Generation for World Complexity]]>https://blog.youkuaiyun.com/qq_41895747/article/details/160209400https://blog.youkuaiyun.com/qq_41895747/article/details/160209400qq_41895747Thu, 16 Apr 2026 11:27:27 +0800<![CDATA[Diffusion Models和MoE：论文、代码汇总]]>https://blog.youkuaiyun.com/qq_41895747/article/details/156332922https://blog.youkuaiyun.com/qq_41895747/article/details/156332922qq_41895747Sat, 27 Dec 2025 13:02:34 +0800<![CDATA[每日AIGC最新进展(94)：MIT提出实时流式视频生成StreamDiffusionV2、LeCun&李飞飞&谢赛宁联合提出空间超感知Cambrian-S]]>https://blog.youkuaiyun.com/qq_41895747/article/details/154752532https://blog.youkuaiyun.com/qq_41895747/article/details/154752532qq_41895747Fri, 14 Nov 2025 08:15:00 +0800<![CDATA[从近期Kimi-Linear、LongCat-Video和Qwen-Next解读下一代大模型架构升级]]>https://blog.youkuaiyun.com/qq_41895747/article/details/154751755https://blog.youkuaiyun.com/qq_41895747/article/details/154751755qq_41895747Thu, 13 Nov 2025 20:30:00 +0800<![CDATA[详细解读视频生成模型Wan2.1代码]]>https://blog.youkuaiyun.com/qq_41895747/article/details/148461132https://blog.youkuaiyun.com/qq_41895747/article/details/148461132qq_41895747Wed, 12 Nov 2025 20:39:43 +0800<![CDATA[每日AIGC最新进展(93)：字节开源720p视频生成自回归模型、字节开源多模态图片编辑模型DreamOmni2、上交提出细粒度PixCLIP]]>https://blog.youkuaiyun.com/qq_41895747/article/details/154664122https://blog.youkuaiyun.com/qq_41895747/article/details/154664122qq_41895747Tue, 11 Nov 2025 08:15:00 +0800<![CDATA[Diffusion Models与视频超分(3): 解读当前最快和最强的开源模型FlashVSR]]>https://blog.youkuaiyun.com/qq_41895747/article/details/154457606https://blog.youkuaiyun.com/qq_41895747/article/details/154457606qq_41895747Mon, 10 Nov 2025 23:10:07 +0800<![CDATA[每日AIGC最新进展(92)：达特茅斯学院提出超高分辨率图片生成Scale-DiT、北航提出滤镜图片生成Personalized Image Filter、清华提出多轮视觉编辑ConsistEdit]]>https://blog.youkuaiyun.com/qq_41895747/article/details/154289045https://blog.youkuaiyun.com/qq_41895747/article/details/154289045qq_41895747Tue, 04 Nov 2025 08:00:00 +0800<![CDATA[每日AIGC最新进展(91)：通义实验室提出Diffusion-MoE方法Routing Matters in MoE、快手提出组相对注意力图像编辑模型GRAG、腾讯提出误差感知趋势一致加速方法ETC]]>https://blog.youkuaiyun.com/qq_41895747/article/details/154288331https://blog.youkuaiyun.com/qq_41895747/article/details/154288331qq_41895747Mon, 03 Nov 2025 08:30:00 +0800<![CDATA[每日AIGC最新进展(90)：腾讯提出视频编辑模型GenCompositor、北京智源研究院发布世界模型Emu3.5]]>https://blog.youkuaiyun.com/qq_41895747/article/details/154285827https://blog.youkuaiyun.com/qq_41895747/article/details/154285827qq_41895747Sun, 02 Nov 2025 14:22:02 +0800<![CDATA[每日AIGC最新进展(89)：Vivo提出大规模超分辨率数据集、希伯来大学提出超高分辨率位置编码外推方法、哈工大提出个性化文生图模型EchoDistill]]>https://blog.youkuaiyun.com/qq_41895747/article/details/154097562https://blog.youkuaiyun.com/qq_41895747/article/details/154097562qq_41895747Thu, 30 Oct 2025 23:30:00 +0800<![CDATA[Diffusion Model与视频超分(2)：解读字节开源视频增强模型SeedVR2]]>https://blog.youkuaiyun.com/qq_41895747/article/details/151905654https://blog.youkuaiyun.com/qq_41895747/article/details/151905654qq_41895747Wed, 29 Oct 2025 23:09:50 +0800<![CDATA[Diffusion Model与视频超分(1)：解读淘宝开源的视频增强模型Vivid-VR]]>https://blog.youkuaiyun.com/qq_41895747/article/details/151904619https://blog.youkuaiyun.com/qq_41895747/article/details/151904619qq_41895747Sat, 20 Sep 2025 22:58:51 +0800<![CDATA[FantasyHSI：Video-Generation-Centric 4D Human Synthesis In Any Scene through A Graph-based ]]>https://blog.youkuaiyun.com/qq_41895747/article/details/151903999https://blog.youkuaiyun.com/qq_41895747/article/details/151903999qq_41895747Sat, 20 Sep 2025 21:09:36 +0800<![CDATA[扩散模型与强化学习(16)：强化学习的10条实战经验（第二版）]]>https://blog.youkuaiyun.com/qq_41895747/article/details/149725941https://blog.youkuaiyun.com/qq_41895747/article/details/149725941qq_41895747Wed, 30 Jul 2025 09:00:00 +0800<![CDATA[扩散模型与强化学习(15)：强化学习的十条实战经验]]>https://blog.youkuaiyun.com/qq_41895747/article/details/149725153https://blog.youkuaiyun.com/qq_41895747/article/details/149725153qq_41895747Tue, 29 Jul 2025 09:00:00 +0800<![CDATA[代码解读：微调Qwen2.5-Omni 实战]]>https://blog.youkuaiyun.com/qq_41895747/article/details/149484021https://blog.youkuaiyun.com/qq_41895747/article/details/149484021qq_41895747Mon, 28 Jul 2025 20:46:33 +0800<![CDATA[畅游Diffusion数字人(31)：FantasyPortrait打造全新面部驱动引擎]]>https://blog.youkuaiyun.com/qq_41895747/article/details/149483814https://blog.youkuaiyun.com/qq_41895747/article/details/149483814qq_41895747Wed, 23 Jul 2025 08:15:00 +0800<![CDATA[扩散模型与强化学习(14)：强化学习中黑客技能的泛化现象与解决方案]]>https://blog.youkuaiyun.com/qq_41895747/article/details/149483391https://blog.youkuaiyun.com/qq_41895747/article/details/149483391qq_41895747Tue, 22 Jul 2025 09:00:00 +0800<![CDATA[扩散模型与强化学习(13)：In-Context Reward Hacking现象与解决方案]]>https://blog.youkuaiyun.com/qq_41895747/article/details/149483164https://blog.youkuaiyun.com/qq_41895747/article/details/149483164qq_41895747Mon, 21 Jul 2025 09:00:00 +0800