•1 min read•from Machine Learning
Predicting human preference for generated image pairs using HPSv3 [P]
Hey! I'm looking for ways to predict human preference for a project I'm building. (imagebench.ai)
I've tryed HPSv3, https://github.com/MizzenAI/HPSv3 and made post about it here:
https://imagebench.ai/blog/does-the-score-match-your-eye
It looks ok, but have many limitation as you can see in my post.
My question. Have you tried other human preference model and found one that would be better then HPSv3?
[link] [comments]
Want to read more?
Check out the full article on the original site
Tagged with
#rows.com
#natural language processing for spreadsheets
#generative AI for data analysis
#Excel alternatives for data analysis
#Human Preference
#HPSv3
#Image Generation
#Image Pairs
#Preference Modeling
#Machine Learning
#Image Evaluation
#AI Models
#Generated Images
#Prediction
#ImageBench
#Model Comparison
#Limitation
#Score Matching
#Reddit
#dh7net