FocusDiT: Masking Queries in Diffusion Transformers for Fine-grained Image Generation
On the map: Diffusion models → Image generation, text-to-image and editing
Abstract (arXiv)
Diffusion transformer (DiT) has been widely adopted in the generative diffusion field, advancing the denoising of query tokens through attention and Feed-Forward (FFN) layers. FFN actually acts as the key-value vocabulary for decoding visual contents where the value embeds the visual semantical knowledge. We present that focusing on critical query tokens corresponding to more complex details and encouraging the model to improve these tokens is essential for fine-grained visual generation. To this end, we propose FocusDiT, which applies a Masking scheme to focus on critical query tokens that are exclusively fed into FFN. The masked queries can retrieve visual tokens from the FFN vocabularies, and use them to decode their visual details. Extensive text-to-image experiments validate the effectiveness of token masking in enhancing generative performance.
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Image generation, text-to-image and editing, one of 14 directions in Diffusion models. Builds on 22 of its 42 references that are in the map. First posted on arXiv 1 Jun 2026.
Builds on
22 of 42 references.
Closest ancestors · 1
Same direction, newest first
Other references · 21
- 2017Attention Is All You Needcited by 60,361
- 2020Denoising Diffusion Implicit Modelscited by 9,781
- 2025Qwen2.5-VL Technical Reportcited by 5,587
- 2024Scaling Rectified Flow Transformers for High-Resolution Image Synthesiscited by 4,181
- 2021Improved Denoising Diffusion Probabilistic Modelscited by 3,761
- 2024Expanding Performance Boundaries of Open-Source Multimodal Models with Model, Data, and Test-Time Scalingcited by 1,599
- 2020Transformer Feed-Forward Layers Are Key-Value Memoriescited by 1,336
- 2023ImageReward: Learning and Evaluating Human Preferences for Text-to-Image Generationcited by 1,045
- 2022Diffusion Models in Vision: A Surveycited by 1,015
- 2019How Can We Know What Language Models Know?cited by 959
- 2021Transformer in Transformercited by 633
- 2021Knowledge Neurons in Pretrained Transformerscited by 601
9 more
- 2022Transformer Feed-Forward Layers Build Predictions by Promoting Concepts in the Vocabulary Spacecited by 532
- 2024PIXART-δ: Fast and Controllable Image Generation with Latent Consistency Modelscited by 141
- 2023DiffRate : Differentiable Compression Rate for Efficient Vision Transformerscited by 77
- 2024Scaling Diffusion Transformers to 16 Billion Parameterscited by 66
- 2024DiG: Scalable and Efficient Diffusion Models with Gated Linear Attentioncited by 56
- 2021BERTnesia: Investigating the capture and forgetting of knowledge in BERTcited by 40
- 2022Kformer: Knowledge Injection in Transformer Feed-Forward Layerscited by 37
- 2021Analyzing Commonsense Emergence in Few-shot Knowledge Modelscited by 34
- 2024EC-DIT: Scaling Diffusion Transformers with Adaptive Expert-Choice Routingcited by 30
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2026Dynamic Differential Linear Attention: Enhancing Linear Diffusion Transformer for High-Quality Image Generation
- 2025UNCAGE: Contrastive Attention Guidance for Masked Generative Transformers in Text-to-Image Generation
- 2025Rare Text Semantics Were Always There in Your Diffusion Transformer
- 2026Learning to Read the Contextual Tokens in Diffusion Transformers
- 2025Muddit: Liberating Generation Beyond Text-to-Image with a Unified Discrete Diffusion Model
- 2024Enhancing Semantic Fidelity in Text-to-Image Synthesis: Attention Regulation in Diffusion Models
- 2026Text Template Tokens Are Implicit Semantic Registers in Diffusion Transformers
- 2024CLEAR: Conv-Like Linearization Revs Pre-Trained Diffusion Transformers Up