Zero-Shot Dynamic Concept Personalization with Grid-Based LoRA

TL;DR

This paper presents a zero-shot framework for dynamic concept personalization in text-to-video generation, enabling high-quality, identity-preserving videos without per-instance fine-tuning by using structured video grids and lightweight adapters.

Contribution

It introduces a novel zero-shot approach using Grid-LoRA adapters and a Grid Fill module for dynamic concept editing in text-to-video models, eliminating the need for test-time optimization.

Findings

High-quality, consistent results across diverse subjects.

Effective generalization to unseen dynamic concepts.

Operates in a single forward pass without fine-tuning.

Abstract

Recent advances in text-to-video generation have enabled high-quality synthesis from text and image prompts. While the personalization of dynamic concepts, which capture subject-specific appearance and motion from a single video, is now feasible, most existing methods require per-instance fine-tuning, limiting scalability. We introduce a fully zero-shot framework for dynamic concept personalization in text-to-video models. Our method leverages structured 2x2 video grids that spatially organize input and output pairs, enabling the training of lightweight Grid-LoRA adapters for editing and composition within these grids. At inference, a dedicated Grid Fill module completes partially observed layouts, producing temporally coherent and identity preserving outputs. Once trained, the entire system operates in a single forward pass, generalizing to previously unseen dynamic concepts without any test-time optimization. Extensive experiments demonstrate high-quality and consistent results across a wide range of subjects beyond trained concepts and editing scenarios.