louisYen/Gen4Gen
🏞️ Official implementation of "Gen4Gen: Generative Data Pipeline for Generative Multi-Concept Composition"
This project helps graphic designers and content creators develop unique visual content by combining specific images of objects with various scene descriptions. It takes individual image assets (like a unique cat or a specific houseplant) and text prompts describing a scene, then generates new, realistic images where these specific objects are naturally placed within complex compositions, complete with detailed text descriptions for each generated image. Anyone needing to create diverse training datasets for personalized image generation models would find this valuable.
110 stars.
Use this if you need to create a large dataset of images featuring your specific assets (e.g., product photos, brand mascots) integrated into a wide variety of generated scenes with accompanying text descriptions.
Not ideal if you're looking for a simple tool to edit existing images or to generate images without providing specific personal assets to incorporate.
Stars
110
Forks
4
Language
Python
License
MIT
Category
Last pushed
Nov 24, 2025
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/louisYen/Gen4Gen"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Related models
AIML-MED/Mirror-CFE
[ICCV25] Looking in the Mirror: A Faithful Counterfactual Explanation Method for Interpreting...
hanzopgp/BrainDataAugmentation
Brain EEG signal data augmentation using synthetic signals generated by a single step diffusion...
anurag-3-nair/Synthetic-Driving-Data-Generation-Pipeline
A personal project to investigate a machine-learning synthetic data generation.