InstantCharacter: Personalize Any Characters with a Scalable Diffusion Transformer Framework

ICLR 2026 Conference Submission18750 Authors

Keywords: Personalized Character Generation; Diffusion Transformer

Abstract: Current learning-based subject customization approaches, predominantly relying on U-Net architectures, suffer from limited generalization ability and compromised image quality. Meanwhile, optimization-based methods require subject-specific fine-tuning, which inevitably degrades textual controllability. To address these challenges, we propose InstantCharacter—a scalable framework for character customization built upon a foundation diffusion transformer. InstantCharacter demonstrates three fundamental advantages: first, it achieves open-domain personalization across diverse character appearances, poses, and styles while maintaining high-fidelity results. Second, we introduce a scalable dual-adapter architecture with stacked transformer encoders, which effectively processes open-domain character features and seamlessly interacts with the latent space of modern diffusion transformers. Third, to effectively train the framework, we construct a large-scale character dataset containing 10-million-level samples. The dataset is systematically organized into paired (multi-view character) and unpaired (text-image combinations) subsets. Our dual-adapter structure addresses the challenge of generating multi-character images by enhancing subject consistency through the image adapter and improving layout control of multiple subjects through the text adapter. Qualitative experiments demonstrate the advanced capabilities of InstantCharacter in generating high-fidelity, text-controllable, and character-consistent images, setting a new benchmark for character-driven image generation.

Primary Area: generative models

Submission Number: 18750