LiON-LoRA: Rethinking LoRA Fusion to Unify Controllable Spatial and Temporal Generation for Video Diffusion

• URL: http://arxiv.org/abs/2507.05678v1
• Date: Tue, 08 Jul 2025 05:00:17 GMT
• Title: LiON-LoRA: Rethinking LoRA Fusion to Unify Controllable Spatial and Temporal Generation for Video Diffusion
• Authors: Yisu Zhang, Chenjie Cao, Chaohui Yu, Jianke Zhu,
• Abstract summary: Video Diffusion Models (VDMs) have demonstrated remarkable capabilities in synthesizing realistic videos by learning from large-scale data.<n>We propose LiON-LoRA, a novel framework that rethinks LoRA fusion through three core principles: Linear scalability, Orthogonality, and Norm consistency.<n> Experiments demonstrate that LiON-LoRA outperforms state-of-the-art methods in trajectory control accuracy and motion strength adjustment, achieving superior generalization with minimal training data.
• Score: 20.022547219190013
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Video Diffusion Models (VDMs) have demonstrated remarkable capabilities in synthesizing realistic videos by learning from large-scale data. Although vanilla Low-Rank Adaptation (LoRA) can learn specific spatial or temporal movement to driven VDMs with constrained data, achieving precise control over both camera trajectories and object motion remains challenging due to the unstable fusion and non-linear scalability. To address these issues, we propose LiON-LoRA, a novel framework that rethinks LoRA fusion through three core principles: Linear scalability, Orthogonality, and Norm consistency. First, we analyze the orthogonality of LoRA features in shallow VDM layers, enabling decoupled low-level controllability. Second, norm consistency is enforced across layers to stabilize fusion during complex camera motion combinations. Third, a controllable token is integrated into the diffusion transformer (DiT) to linearly adjust motion amplitudes for both cameras and objects with a modified self-attention mechanism to ensure decoupled control. Additionally, we extend LiON-LoRA to temporal generation by leveraging static-camera videos, unifying spatial and temporal controllability. Experiments demonstrate that LiON-LoRA outperforms state-of-the-art methods in trajectory control accuracy and motion strength adjustment, achieving superior generalization with minimal training data. Project Page: https://fuchengsu.github.io/lionlora.github.io/

Related papers

• Rethinking LoRA for Privacy-Preserving Federated Learning in Large Models [14.755143405057929]
  Fine-tuning large vision models (LVMs) and large language models (LLMs) under differentially private learning (DPFL) is hindered by a fundamental privacy-utility trade-off.<n>Low-Rank Adaptation (LoRA), a promising parameter-efficient fine-tuning (PEFT) method, reduces computational and communication costs by introducing two trainable low-rank matrices while freezing pre-trained weights.<n>We propose LA-LoRA, a novel approach that decouples gradient interactions and aligns update directions across clients to enhance robustness under stringent privacy constraints.
  arXiv  Detail & Related papers  (2026-02-23T15:05:28Z)
• Driving with DINO: Vision Foundation Features as a Unified Bridge for Sim-to-Real Generation in Autonomous Driving [36.98878302668877]
  We present Driving with DINO (DwD), a novel framework for autonomous driving video generation.<n>We first identify that these features encode a spectrum of information, from high-level semantics to fine-grained structure.<n>To effectively utilize this, we employ Principal Subspace Projection to discard the high-frequency elements responsible for "texture baking"
  arXiv  Detail & Related papers  (2026-02-05T19:55:22Z)
• NP-LoRA: Null Space Projection Unifies Subject and Style in LoRA Fusion [16.64405108290577]
  Low-Rank Adaptation (LoRA) fusion is a key technique for reusing and composing learned subject and style representations.<n>Existing methods rely on weight-based merging, where one LoRA often dominates the other.<n>We propose Null Space Projection LoRA, a projection-based framework for LoRA fusion that enforces subspace separation.
  arXiv  Detail & Related papers  (2025-11-14T08:06:01Z)
• ScaleWeaver: Weaving Efficient Controllable T2I Generation with Multi-Scale Reference Attention [86.93601565563954]
  ScaleWeaver is a framework designed to achieve high-fidelity, controllable generation upon advanced visual autoregressive( VAR) models.<n>The proposed Reference Attention module discards the unnecessary attention from image$rightarrow$condition, reducing computational cost.<n>Experiments show that ScaleWeaver delivers high-quality generation and precise control while attaining superior efficiency over diffusion-based methods.
  arXiv  Detail & Related papers  (2025-10-16T17:00:59Z)
• Real-Time Motion-Controllable Autoregressive Video Diffusion [79.32730467857535]
  We propose AR-Drag, the first RL-enhanced few-step AR video diffusion model for real-time image-to-video generation with diverse motion control.<n>We first fine-tune a base I2V model to support basic motion control, then further improve it via reinforcement with a trajectory-based reward model.<n>Our design preserves the Markov property through a Self-Rollout learning mechanism and accelerates training by selectively denoising steps.
  arXiv  Detail & Related papers  (2025-10-09T12:17:11Z)
• AutoLoRA: Automatic LoRA Retrieval and Fine-Grained Gated Fusion for Text-to-Image Generation [32.46570968627392]
  Low-rank adaptation (LoRA) have demonstrated efficacy in enabling model customization with minimal parameter overhead.<n>We introduce a novel framework that enables semantic-driven LoRA retrieval and dynamic aggregation.<n>Our approach achieves significant improvement in image generation perfermance.
  arXiv  Detail & Related papers  (2025-08-04T06:36:00Z)
• One-Step Diffusion for Detail-Rich and Temporally Consistent Video Super-Resolution [9.03810927740921]
  We propose a Dual LoRA Learning (DLoRAL) paradigm to train an effective SD-based one-step diffusion model.<n> Experiments show that DLoRAL achieves strong performance in both accuracy and speed.
  arXiv  Detail & Related papers  (2025-06-18T16:06:30Z)
• DragLoRA: Online Optimization of LoRA Adapters for Drag-based Image Editing in Diffusion Model [14.144755955903634]
  DragLoRA is a novel framework that integrates LoRA adapters into the drag-based editing pipeline.<n>We show that DragLoRA significantly enhances the control precision and computational efficiency for drag-based image editing.
  arXiv  Detail & Related papers  (2025-05-18T13:52:19Z)
• SuperFlow++: Enhanced Spatiotemporal Consistency for Cross-Modal Data Pretraining [62.433137130087445]
  SuperFlow++ is a novel framework that integrates pretraining and downstream tasks using consecutive camera pairs.<n>We show that SuperFlow++ outperforms state-of-the-art methods across diverse tasks and driving conditions.<n>With strong generalizability and computational efficiency, SuperFlow++ establishes a new benchmark for data-efficient LiDAR-based perception in autonomous driving.
  arXiv  Detail & Related papers  (2025-03-25T17:59:57Z)
• Bench2Drive-R: Turning Real World Data into Reactive Closed-Loop Autonomous Driving Benchmark by Generative Model [63.336123527432136]
  We introduce Bench2Drive-R, a generative framework that enables reactive closed-loop evaluation.<n>Unlike existing video generative models for autonomous driving, the proposed designs are tailored for interactive simulation.<n>We compare the generation quality of Bench2Drive-R with existing generative models and achieve state-of-the-art performance.
  arXiv  Detail & Related papers  (2024-12-11T06:35:18Z)
• Dual Low-Rank Adaptation for Continual Learning with Pre-Trained Models [38.97142043836567]
  Continual learning (CL) aims to enable vision transformers (ViTs) to learn new tasks over time.<n> catastrophic forgetting remains a persistent challenge.<n>We propose a novel PEFT-CL method called Dual Low-Rank Adaptation (DualLoRA)
  arXiv  Detail & Related papers  (2024-11-01T14:28:39Z)
• LCM-LoRA: A Universal Stable-Diffusion Acceleration Module [52.8517132452467]
  Latent Consistency Models (LCMs) have achieved impressive performance in accelerating text-to-image generative tasks. This report further extends LCMs' potential by applying LoRA distillation to larger Stable-Diffusion models. We identify the LoRA parameters obtained through LCM distillation as a universal Stable-Diffusion acceleration module, named LCM-LoRA.
  arXiv  Detail & Related papers  (2023-11-09T18:04:15Z)
• Interactive Character Control with Auto-Regressive Motion Diffusion Models [18.727066177880708]
  We propose A-MDM (Auto-regressive Motion Diffusion Model) for real-time motion synthesis. Our conditional diffusion model takes an initial pose as input, and auto-regressively generates successive motion frames conditioned on previous frame. We introduce a suite of techniques for incorporating interactive controls into A-MDM, such as task-oriented sampling, in-painting, and hierarchical reinforcement learning.
  arXiv  Detail & Related papers  (2023-06-01T07:48:34Z)
• Benchmarking the Robustness of LiDAR-Camera Fusion for 3D Object Detection [58.81316192862618]
  Two critical sensors for 3D perception in autonomous driving are the camera and the LiDAR. fusing these two modalities can significantly boost the performance of 3D perception models. We benchmark the state-of-the-art fusion methods for the first time.
  arXiv  Detail & Related papers  (2022-05-30T09:35:37Z)
• TransFusion: Robust LiDAR-Camera Fusion for 3D Object Detection with Transformers [49.689566246504356]
  We propose TransFusion, a robust solution to LiDAR-camera fusion with a soft-association mechanism to handle inferior image conditions. TransFusion achieves state-of-the-art performance on large-scale datasets. We extend the proposed method to the 3D tracking task and achieve the 1st place in the leaderboard of nuScenes tracking.
  arXiv  Detail & Related papers  (2022-03-22T07:15:13Z)
• LIF-Seg: LiDAR and Camera Image Fusion for 3D LiDAR Semantic Segmentation [78.74202673902303]
  We propose a coarse-tofine LiDAR and camera fusion-based network (termed as LIF-Seg) for LiDAR segmentation. The proposed method fully utilizes the contextual information of images and introduces a simple but effective early-fusion strategy. The cooperation of these two components leads to the success of the effective camera-LiDAR fusion.
  arXiv  Detail & Related papers  (2021-08-17T08:53:11Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.