Not my model, from the huggingface repo. This is an excellent merge model, particularly in the middle blocks. Try it yourself - take your favorite model, and block merge this at about 10% input, and 20% middle, and adjust from there.Original U-Net: <a target="_blank" rel="ugc" href="https://huggingface.co/mhdang/dpo-sd1.5-text2image-v1">https://huggingface.co/mhdang/dpo-sd1.5-text2image-v1</a>bdsqlz's release: <a target="_blank" rel="ugc" href="https://huggingface.co/bdsqlsz/dpo-sd-text2image-v1-fp16">https://huggingface.co/bdsqlsz/dpo-sd-text2image-v1-fp16</a>bdsqlz released the sdxl model here: <a target="_blank" rel="ugc" href="https://civarchive.com/models/237681/dpo-sdxl-fp16">https://civarchive.com/models/237681/dpo-sdxl-fp16</a> but us poor 1.5 users were left in the dark ages.I had to do some hacking to get the fp32 version, so you will have to bring your own VAE.<h1 id="heading-62">Diffusion Model Alignment Using Direct Preference Optimization</h1>Direct Preference Optimization (DPO) for text-to-image diffusion models is a method to align diffusion models to text human preferences by directly optimizing on human comparison data. Please check paper at <a target="_blank" rel="ugc" href="https://arxiv.org/abs/2311.12908">Diffusion Model Alignment Using Direct Preference Optimization</a>.SD1.5 model is fine-tuned from <a target="_blank" rel="ugc" href="https://huggingface.co/runwayml/stable-diffusion-v1-5">stable-diffusion-v1-5</a> on offline human preference data <a target="_blank" rel="ugc" href="https://huggingface.co/datasets/yuvalkirstain/pickapic_v2">pickapic_v2</a>.SDXL model is fine-tuned from <a target="_blank" rel="ugc" href="https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0">stable-diffusion-xl-base-1.0</a> on offline human preference data <a target="_blank" rel="ugc" href="https://huggingface.co/datasets/yuvalkirstain/pickapic_v2">pickapic_v2</a>.

Files with SHA256: e349bc2c334e6ce0fe9e985469f3396da0af9e6592b9856646483b1f8a13e8d1