AI Paper Atlas

Research map / Image generation / Text-to-image diffusion

ZipLoRA: Any Subject in Any Style by Effectively Merging LoRAs

Viraj Shah, Nataniel Ruiz, Forrester Cole, Erika Lu, Svetlana Lazebnik, Yuanzhen Li, Varun Jampani

ECCV 2024 · Image generation › Text-to-image diffusion · approach: Multi-subject personalization · 149 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
merge LoRAs with learned per-column coefficients and cosine-similarity penalty instead of direct arithmetic sumZipLoRA · Subject-style mergingsubject-alignment (DINO): 0.357 -> 0.420Improved
merge LoRAs with ZipLoRA instead of direct arithmetic sumZipLoRA · Subject-style mergingstyle-alignment (CLIP-I): 0.702 -> 0.699No clear effect
use ZipLoRA instead of direct merge, judged by usersZipLoRA · User study% preference: 82.7% prefer ZipLoRAImproved
use ZipLoRA instead of joint trainingZipLoRA · Runtimeseconds: 3540 -> 560Improved
optimize only merger coefficient vectors with frozen LoRAsZipLoRA · EfficiencyGPU memory / storage: 38GB -> 21GB; 360MB -> 6.5MBImproved
use SDXL as base instead of SD v1.5 for style LoRAs (ablation)ZipLoRA · Style learning from one imagestylization qualityImproved
apply ZipLoRA to styles whose reference image contains content such as a cliffZipLoRA · Failure casescontent leakageHurt
zero out 80-90% of lowest-magnitude LoRA update elements (ablation)LoRA update matrices · Qualitative (watercolor teddy bear)generation qualityNo clear effect

Later papers that compared with this method

5 later accepted papers report ZipLoRA as a baseline in their results tables; 4 of them report a better number on most of the results they share with it.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy