← Back to all news
arXiv Machine Learning By Minkyu Kim, Baekseung Kim, Junhoo Lee, Jangho Kim, Nojun Kwak

ManiFusion: Unlocking the High-Throughput Generation via Superposition in Manifold Space

Read the original on arXiv Machine Learning →

ManiFusion is a backbone‑agnostic framework that boosts the throughput of diffusion models by fusing multiple samples into a single denoising evaluation and then unfusing them afterward. It introduces Trajectory‑Fusion Scheduling to dynamically adjust fusion factors across sampling stages, allowing shared representations while preserving individual generation paths. The method works with various diffusion families, including U‑Net and transformer backbones in pixel and latent spaces, and can be combined with existing acceleration techniques to achieve higher throughput with minimal quality loss.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.

Related stories

arXiv Computer Vision
Oct 12

InstantFusion: A Shared Latent Interface for Heterogeneous Image Generators

InstantFusion introduces a lightweight shared latent interface that enables cross-model state translation between heterogeneous image generators. By training model-specific Latent AutoEncoders to project intermediate denoising states into a common latent space, the method aligns states from the same image at matched noise levels and synchronizes their local denoising velocities. This allows seamless denoising transitions across frozen generators, supporting applications such as acceleration, multi-preference composition, on-policy distillation, and cross-model composition.

By Xuanlang Dai, Jiazi Bu, Yujie Zhou, Yuhong Liu, Beichen Zhang, Penghui Yang, Ziyu Liu, Yuhang Zang
diffusionefficiencysafety
More like this →
arXiv Computer Vision
Aug 24

RecGen3D: Reconstruction-Guided 3D Generation in a Shared Canonical Space

RecGen3D is a framework that merges feed‑forward reconstruction and diffusion‑based generation to address the trade‑off between reconstruction fidelity and generative plausibility in sparse‑view 3D modeling. By aligning both models in a shared canonical space and using decoupled cooperative learning, the system stabilizes training and allows the reconstruction module to supply canonical geometric anchors while the diffusion generator refines and completes the structure. Experiments show that RecGen3D outperforms existing methods in producing complete and consistent 3D models from sparse observations.

By Zhisheng Huang, Jiahao Chen, Cheng Lin, Chenyu Hu, Hanzhuo Huang, Zhengming Yu, Mengfei Li, Yuheng Liu, Zekai Gu, Zibo Zhao, Yuan Liu, Xin Li, Wenping Wang
diffusionsafety
More like this →