<h1 id="this-is-a-lora-oriented-on-dall-e-3-compositions.-795i8354r">This is a Lora oriented on Dall-E 3 compositions.</h1>Main trigger words : D4lle, Dall-e style, AI generated////////////////////Datasets infos: 30 pictures (Humans images)<ul><li>80% solo / 20% duo</li><li>74% woman / 26% man</li><li>50% conceptual, themed, poetic aesthetic / 50% scenes (various contexts)</li></ul>TIPS: Also useful for fixing bad renderings generated on Dall-E, using this Lora in img2img allows you to best preserve the initial Dall-E style.____________________<h2 id="if-you-want-to-see-my-next-lora-projects-(for-flux)-and-contribute-to-launch-the-model-training-:-7b0757qnb">If you want to see my next Lora projects (for Flux) and contribute to launch the model training :</h2>Donations for futurs loras : <a target="_blank" rel="ugc" href="https://ko-fi.com/nicolas1801">https://ko-fi.com/nicolas1801</a>Some details: <a target="_blank" rel="ugc" href="https://docs.google.com/spreadsheets/d/1PEmIlzhUUGeoPZqRafKOnvxNzR_-a3SH/edit?usp=sharing&amp;ouid=104542247627018372519&amp;rtpof=true&amp;sd=true">https://docs.google.com/spreadsheets/d/1PEmIlzhUUGeoPZqRafKOnvxNzR_-a3SH/edit?usp=sharing&amp;ouid=104542247627018372519&amp;rtpof=true&amp;sd=true</a>

Dall-E 3 / Bing - Selection 1

Based on a selection of 30 images, so this is the serie 1.2 (RAW captionned with Phi3 128k + WD-vit large v3, not re-edited).

Files with SHA256: c5dedbcfbf6f973359efb681cf42939b81ea020ca08587726638827f31d6119e