Paper page - Lens: Rethinking Training Efficiency for Foundational Text-to-Image Models

Source: https://huggingface.co/papers/2605.21573 Published on May 20

#1 Paper of the day Authors:

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

Abstract

Lens is a compact 3.8B-parameter text-to-image model achieving superior performance with reduced training compute through dense caption datasets, multi-resolution batching, efficient architecture, and optimization techniques.

We introduce Lens, a 3.8B-parameterT2I modelthat achieves performance competitive with, and in several cases surpassing, state-of-the-art models with more than 6B parameters across various benchmarks, while requiring significantly less training compute. For example, Lens requires only about 19.3% of the training compute used by Z-Image. The training efficiency of Lens stems from two key strategies beyond its compact model size. First, we maximize data information density per training batch by (i) training on Lens-800M, a dataset of 800M densely captioned image-text pairs whose captions are generated by GPT-4.1 and contain approximately 109 words on average, providing richer semantic supervision than conventional short captions, and (ii) constructing each batch from images with multiple resolutions and diverse aspect ratios, thereby enlarging the effective visual coverage of each optimization step. Second, we improve convergence speed through careful architectural choices, including adopting a semanticVAEthat provides better latent representations and employing a stronglanguage encoderthat accelerates optimization while enabling multilingual generalization from English-only training data. After pre-training, we applyRLwithtaxonomy-driven prompts(Lens-RL-8K) andstructured reward rubricsto suppress artifacts and improve visual quality, areasoner modulewith training-free system prompt search to better align user requests with the model, anddistillation-based accelerationfor 4-step inference. Through efficient training and systematic optimization, Lens generalizes to arbitrary aspect ratios from 1:2 to 2:1 and resolutions up to 1440^2, and supports prompts in several commonly used languages. Thanks to its compact size, Lens generates a 1024^2 image in 3.15 seconds on a single NVIDIA H100 GPU, while its distilled turbo version performs 4-step generation in 0.84 seconds.

View arXiv pageView PDFProject pageGitHubAdd to collection

Get this paper in your agent:

hf papers read 2605\.21573

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper4

#### microsoft/Lens-Turbo Text-to-Image• Updated3 days ago • 543 • 72 #### microsoft/Lens Text-to-Image• Updated3 days ago • 434 • 68 #### microsoft/Lens-Base Text-to-Image• Updated3 days ago • 154 • 13 #### ngoctham/Lens Text-to-Image• Updated2 days ago • 13 • 1

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2605.21573 in a dataset README.md to link it from this page.

Spaces citing this paper1

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.