enfugue/DPO (Direct Preference Optimization) LoRA for XL and 1.5 - OpenRail++
What is DPO? DPO is D irect P reference O ptimization, the name given to the process whereby a diffusion model is finetuned based on human-chosen images. Meihua Dang et. al. have trained Stable Diffusion 1.5 and Stable Diffusion XL using this method and the Pick-a-Pic v2 Dataset, which can be found at https://huggingface.co/datasets/yuvalkirstain/pickapic_v2 , and wrote a paper about it at https://huggingface.co/papers/2311.12908 . What does it Do? The trained DPO models have been observed to produce higher quality images than their untuned counterparts, with a significant emphasis on the adhe
| Author / organization | enfugue |
|---|---|
| Downloads | 18.3K (as reported by Civitai) |
| Likes | 1.4K |
| Task | text-to-image |
| Created | 24 December 2023 |
| Last updated at source | — |
Popularity, last 2 days
18,262 now (+5 since 2026-10-15). Counts as reported by the source, snapshotted daily by Silytics.
Silytics shows public metadata from Civitai and links to the original. Information may be incomplete or out of date. Last refreshed 16 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.