DreamLite enables sub-second 1024x1024 image generation and editing on mobile devices using a unified 0.39B parameter model.
March 31, 2026
Original Paper
DreamLite: A Lightweight On-Device Unified Model for Image Generation and Editing
arXiv · 2603.28713
The Takeaway
This is the first on-device model to unify high-resolution text-to-image generation and text-guided editing in a single lightweight network. It bridges the gap between server-side model capabilities and the strict latency/memory constraints of edge deployment on smartphones.
From the abstract
Diffusion models have made significant progress in both text-to-image (T2I) generation and text-guided image editing. However, these models are typically built with billions of parameters, leading to high latency and increased deployment challenges. While on-device diffusion models improve efficiency, they largely focus on T2I generation and lack support for image editing. In this paper, we propose DreamLite, a compact unified on-device diffusion model (0.39B) that supports both T2I generation a