Research map / Image generation / Text-to-image diffusion
MakeAnything: Harnessing Diffusion Transformers for Multi-Domain Procedural Sequence Generation
Yiren Song, Cheng Liu, Mike Zheng Shou
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 changed | Where | Measured | Effect |
|---|---|---|---|
| use shared-A asymmetric LoRA instead of mixed-data standard LoRA (ablation) | MakeAnything · avg over 21 tasks | GPT alignment: 3.72 -> 4.35 | Improved |
| use shared-A asymmetric LoRA instead of mixed-data standard LoRA (ablation) | MakeAnything · avg over 21 tasks | GPT coherence: 3.50 -> 3.97 | Improved |
| use shared-A asymmetric LoRA instead of single-task standard LoRA (ablation) | MakeAnything · avg over 21 tasks | GPT alignment: 3.68 -> 4.35 | Improved |
| use shared-A asymmetric LoRA instead of single-task standard LoRA (ablation) | MakeAnything · avg over 21 tasks | CLIP score: 29.42 -> 32.70 | Improved |
| use shared-A asymmetric LoRA instead of single-task standard LoRA (ablation) | MakeAnything · avg over 21 tasks | GPT usability: 3.76 -> 4.09 | Improved |
| use mixed-data standard LoRA versus single-task LoRA (ablation) | MakeAnything · avg over 21 tasks | GPT coherence: 3.83 -> 3.50 | Hurt |
| MakeAnything versus baselines (Flux, Ideogram, ProcessPainter) | MakeAnything · text-to-process | GPT/human preference | Improved |
| add serpentine frame layout | MakeAnything · sequence grids | coherence | No clear effect |
Later papers that compared with this method
1 later accepted paper reports MakeAnything as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- ProcessMaker: A Generalized Process Visualization Framework with Adaptive Sequence Steps on Diffusion Transformers (CVPR 2026 · better on 7 of 7 shared results)
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
- Mastering Text-to-Image Diffusion: Recaptioning, Planning, and Generating with Multimodal LLMs (ICML 2024 · similarity 0.86)
- VEDIT: Latent Prediction Architecture For Procedural Video Representation Learning (ICLR 2025 · similarity 0.86)
- Procedural Level Generation with Diffusion Models from a Single Example (AAAI 2024 · similarity 0.85)
- Enabling Instructional Image Editing with In-Context Generation in Large Scale Diffusion Transformer (NeurIPS 2025 · similarity 0.85)
- Generate Subgoal Images before Act: Unlocking the Chain-of-Thought Reasoning in Diffusion Model for Robot Manipulation with Multimodal Prompts (CVPR 2024 · similarity 0.85)
- ACE: All-round Creator and Editor Following Instructions via Diffusion Transformer (ICLR 2025 · similarity 0.85)
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.
