从 Peebles 与谢赛宁的 DiT 论文到 Sora 的时空 patch，讲清 Diffusion Transformer 的三个关键设计：patch 化、adaLN-Zero 条件注入与以计算量为标尺的可扩展性。

盘点 2026 年 10 月主流开源 TTS 七个项目（GPT-SoVITS、CosyVoice、F5-TTS、Fish Speech、IndexTTS、Kokoro 等）：机制、音色克隆方式、中文支持与许可证商用限制，附中文效果/实时率/长文本对比表与 F5-TTS 上手示例，兼谈声音克隆的授权与深度伪造合规风险。

模仿学习的老问题是「多峰动作分布」——两种都对的做法被回归平均成一种错的。Diffusion Policy 用条件去噪扩散直接建模动作分布，在 12 个任务、4 个基准上平均成功率提升 46.9%，此后 DP3、DPPO 相继跟进，π0 的流匹配与 GR00T 的扩散 Transformer 把它推成了 VLA 时代的标配动作头。

从 RT-2 把机器人动作当文本 token 输出，到 OpenVLA 开源 7B、π0 用流匹配跑到 50Hz、GR00T 与 Helix 转向双系统架构——本文梳理 VLA 模型三年的演进脉络：动作怎么表示、参数怎么变小、控制频率怎么上去，以及开源与闭源两条路线的分野。

生成一条好看的视频已经不难，难的是让同一个人跨镜头不「变脸」、让画面里的世界遵守物理。盘点各家的参考生视频、LoRA 微调、推理优先架构等方案与仍然存在的短板。

Interactive video world models need to generate each video chunk efficiently while responding faithfully to user control…

Multimodal Diffusion Transformers (MM-DiTs) jointly process visual and textual representations throughout generation. Th…

Native 3D texture generation synthesizes colors directly in 3D space for a given geometry, conditioned on multi-view ref…

Diffusion Transformers (DiTs) have established themselves as a scalable backbone for high-fidelity image synthesis. Howe…

Representation autoencoders (RAEs) reuse features from a pretrained visual encoder as reconstruction and diffusion laten…

思邈* 2026-09-24 22:17:48 来源：量子位 1.5台6000D跑赢1台B300！ 允中 发自 凹非寺 量子位 | 公众号QbitAI 随着大模型对算力需求持续攀升，**GPU卡的采购成本、供给约束持续加剧。** AI推理…

Alibaba's Qwen team has released Qwen-Image-2.1, a 7B diffusion transformer that handles text-to-image generation, multi…

High-quality texture generation is essential for creating realistic and production-ready 3D assets. Recent multi-view di…

Video diffusion models repeatedly process long spatiotemporal token sequences during denoising, making attention a major…

Visual signals require compact yet sufficient representations for robust downstream prediction. Convolutional sparse cod…

We introduce StepAudio 3 Music, a large-scale, long-form music generation model that supports explicit musical planning …

15秒视频，50秒出片]

Monocular depth estimation is a ubiquitous yet highly ill-posed computer vision task, with downstream applications in sc…