# Data Warmup: Complexity-Aware Curricula for Efficient Diffusion Training

arXiv: https://arxiv.org/abs/2604.07397 · HTML: https://mlsanity.com/paper/2604.07397/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 28 of its 39 references that are in the map.
- First posted on arXiv 8 Apr 2026.

## Builds on (most cited first, 28 of 28)

- [Denoising Diffusion Probabilistic Models](https://mlsanity.com/paper/2006.11239/) (2020)
- [High-Resolution Image Synthesis with Latent Diffusion Models](https://mlsanity.com/paper/2112.10752/) (2021)
- [Score-Based Generative Modeling through Stochastic Differential Equations](https://mlsanity.com/paper/2011.13456/) (2020)
- [ImageNet Large Scale Visual Recognition Challenge](https://mlsanity.com/paper/1409.0575/) (2014)
- [Diffusion Models Beat GANs on Image Synthesis](https://mlsanity.com/paper/2105.05233/) (2021)
- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [Masked Autoencoders Are Scalable Vision Learners](https://mlsanity.com/paper/2111.06377/) (2021)
- [Photorealistic Text-to-Image Diffusion Models with Deep Language Understanding](https://mlsanity.com/paper/2205.11487/) (2022)
- [Improved Techniques for Training GANs](https://mlsanity.com/paper/1606.03498/) (2016)
- [Prioritized Experience Replay](https://mlsanity.com/paper/1511.05952/) (2015)
- [Improved Precision and Recall Metric for Assessing Generative Models](https://mlsanity.com/paper/1904.06991/) (2019)
- [SiT: Exploring Flow and Diffusion-based Generative Models with Scalable Interpolant Transformers](https://mlsanity.com/paper/2401.08740/) (2024)
- [Representation Alignment for Generation: Training Diffusion Transformers Is Easier Than You Think](https://mlsanity.com/paper/2410.06940/) (2024)
- [Curriculum Learning for Reinforcement Learning Domains: A Framework and Survey](https://mlsanity.com/paper/2003.04960/) (2020)
- [Not All Samples Are Created Equal: Deep Learning with Importance Sampling](https://mlsanity.com/paper/1803.00942/) (2018)
- [GRAD-MATCH: Gradient Matching based Data Subset Selection for Efficient Deep Model Training](https://mlsanity.com/paper/2103.00123/) (2021)
- [Generating Images with Sparse Representations](https://mlsanity.com/paper/2103.03841/) (2021)
- [Automated Curriculum Learning for Neural Networks](https://mlsanity.com/paper/1704.03003/) (2017)
- [Prioritized Training on Points that are Learnable, Worth Learning, and Not Yet Learnt](https://mlsanity.com/paper/2206.07137/) (2022)
- [Competence-based Curriculum Learning for Neural Machine Translation](https://mlsanity.com/paper/1903.09848/) (2019)
- [Fast Training of Diffusion Models with Masked Transformers](https://mlsanity.com/paper/2306.09305/) (2023)
- [Online Batch Selection for Faster Training of Neural Networks](https://mlsanity.com/paper/1511.06343/) (2015)
- [FasterDiT: Towards Faster Diffusion Transformers Training without Architecture Modification](https://mlsanity.com/paper/2410.10356/) (2024)
- [FiTv2: Scalable and Improved Flexible Vision Transformer for Diffusion Model](https://mlsanity.com/paper/2410.13925/) (2024)
- [SARA: Structural and Adversarial Representation Alignment for Training-efficient Diffusion Models](https://mlsanity.com/paper/2503.08253/) (2025)
- [Efficient Generative Model Training via Embedded Representation Warmup](https://mlsanity.com/paper/2504.10188/) (2025)
- [From Prototypes to General Distributions: An Efficient Curriculum for Masked Image Modeling](https://mlsanity.com/paper/2411.10685/) (2024)
- [Translation from Wearable PPG to 12-Lead ECG](https://mlsanity.com/paper/2509.25480/) (2025)

## Abstract (arXiv)

A key inefficiency in diffusion training occurs when a randomly initialized network, lacking visual priors, encounters gradients from the full complexity spectrum--most of which it lacks the capacity to resolve. We propose Data Warmup, a curriculum strategy that schedules training images from simple to complex without modifying the model or loss. Each image is scored offline by a semantic-aware complexity metric combining foreground dominance (how much of the image salient objects occupy) and foreground typicality (how closely the salient content matches learned visual prototypes). A temperature-controlled sampler then prioritizes low-complexity images early and anneals toward uniform sampling. On ImageNet 256x256 with SiT backbones (S/2 to XL/2), Data Warmup improves IS by up to 6.11 and FID by up to 3.41, reaching baseline quality tens of thousands of iterations earlier. Reversing the curriculum (exposing hard images first) degrades performance below the uniform baseline, confirming that the simple-to-complex ordering itself drives the gains. The method combines with orthogonal accelerators such as REPA and requires only ~10 minutes of one-time preprocessing with zero per-iteration overhead.