shunk031/training-free-structured-diffusion-guidance
🤗 Unofficial huggingface/diffusers-based implementation of the paper "Training-Free Structured Diffusion Guidance for Compositional Text-to-Image Synthesis".
This tool helps digital artists, content creators, and marketers generate images from text descriptions with greater accuracy, especially when multiple objects or specific attributes are involved. You provide a detailed text prompt describing your desired image, and it produces an image where elements like colors, objects, and their relationships are correctly depicted, avoiding common errors like misplaced attributes or missing items.
120 stars. No commits in the last 6 months.
Use this if you need to create images from complex text prompts and frequently encounter issues where attributes (like colors or characteristics) are wrongly applied to objects, or objects are missing from your generated image.
Not ideal if your text-to-image needs are simple and you don't commonly face problems with compositional accuracy, as it adds a layer of guidance that might not be necessary for basic prompts.
Stars
120
Forks
5
Language
Python
License
Apache-2.0
Category
Last pushed
Mar 29, 2023
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/shunk031/training-free-structured-diffusion-guidance"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
OPTML-Group/Unlearn-Saliency
[ICLR24 (Spotlight)] "SalUn: Empowering Machine Unlearning via Gradient-based Weight Saliency in...
Shilin-LU/VINE
[ICLR 2025] "Robust Watermarking Using Generative Priors Against Image Editing: From...
WindVChen/DiffAttack
An unrestricted attack based on diffusion models that can achieve both good transferability and...
koninik/DiffusionPen
Official PyTorch Implementation of "DiffusionPen: Towards Controlling the Style of Handwritten...
Wuyxin/DISC
(ICML 2023) Discover and Cure: Concept-aware Mitigation of Spurious Correlation