AI Paper Atlas

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

NanoFLUX: Distillation-Driven Compression of Large Text-to-Image Generation Models for Mobile Devices

Ruchika Chavhan, Malcolm Chadwick, Alberto Gil Couto Pimentel Ramos, Luca Morreale, Mehdi Noroozi, Abhinav Mehrotra

ICML 2026 · Image generation › Text-to-image diffusion · approach: Layout-to-image control · 0 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
prune heads, 12B to 3B DiT (step C2)FLUX.1-Schnell · GenEvalscore: 66.0 -> 54.1Hurt
compress DiT from 12B to 1.8BFLUX.1-Schnell · One-IGscore: 43.1 -> 43.2No clear effect
add progressive token downsampling to 2.5B DiT (ablation)Step C3 2.5B · HPSv3score: 10.68 -> 10.56Hurt
add progressive token downsampling to 1.8B DiT (ablation)Step C4 1.8B · GenEvalscore: 53.1 -> 49.8Hurt
add progressive token downsampling to 1.8B DiT (ablation)Step C4 1.8B · latency on Samsung S25U (s)seconds: 2.75 -> 2.40Improved
distill T5-XXL into T5-Large text encoderStep C3 2.5B · DPGscore: 82.6 -> 76.3Hurt
distill T5-XXL into T5-Large text encoderStep C3 2.5B · HPSv3score: 10.68 -> 10.45Hurt
replace pruning with block merging in single-stream blocks 7-12 (ablation)2.5B DiT · HPSv3score: 9.66 -> 10.66Improved
prune double-stream blocks instead of single-stream blocks (ablation)2.5B DiT · HPSv3score: 9.66 -> 9.41Hurt
add static AdaLN coefficients precomputed from 2 samples (ablation)2.5B DiT · HPSv3score: 10.66 -> 10.54No clear effect
use block-wise distillation loss instead of output-only distillation (ablation)T5-Large for 2.5B DiT · HPSv3score: 10.37 -> 10.45Improved
use block-wise distillation loss instead of output-only distillation (ablation)T5-Large for 1.8B DiT · HPSv3score: 10.21 -> 10.45Improved

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