mihirp1998/AlignProp
AlignProp uses direct reward backpropogation for the alignment of large-scale text-to-image diffusion models. Our method is 25x more sample and compute efficient than reinforcement learning methods (PPO) for finetuning Stable Diffusion
This project helps AI developers fine-tune large text-to-image models like Stable Diffusion to produce images that better align with specific goals, such as improved aesthetic quality, semantic accuracy, or object controllability. Developers input an existing text-to-image model and a reward function, and the output is a more finely tuned model that generates images optimized for the desired criteria. It's designed for machine learning engineers and researchers working on generative AI applications.
314 stars. No commits in the last 6 months.
Use this if you are a machine learning engineer who needs to efficiently improve the quality and control of images generated by large text-to-image diffusion models for specific downstream tasks.
Not ideal if you are an end-user simply looking to generate images without needing to train or fine-tune models yourself.
Stars
314
Forks
11
Language
Python
License
MIT
Category
Last pushed
Nov 01, 2024
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/mihirp1998/AlignProp"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
FlorianFuerrutter/genQC
Generative Quantum Circuits
horseee/DeepCache
[CVPR 2024] DeepCache: Accelerating Diffusion Models for Free
Gen-Verse/MMaDA
MMaDA - Open-Sourced Multimodal Large Diffusion Language Models (dLLMs with block diffusion,...
kuleshov-group/mdlm
[NeurIPS 2024] Simple and Effective Masked Diffusion Language Model
Shark-NLP/DiffuSeq
[ICLR'23] DiffuSeq: Sequence to Sequence Text Generation with Diffusion Models