These very talented artists permitted me to use their fantastic images. @theendlessimagination <a target="_blank" rel="ugc" href="https://civarchive.com/images/106512005">https://civarchive.com/images/106512005</a> @NeroOnFire <a target="_blank" rel="ugc" href="https://civarchive.com/images/106481390">https://civarchive.com/images/106481390</a> @Mordab <a target="_blank" rel="ugc" href="https://civarchive.com/images/106511996">https://civarchive.com/images/106511996</a> @ecaj <a target="_blank" rel="ugc" href="https://civarchive.com/user/ecaj">https://civarchive.com/user/ecaj</a> The Elsa image he posted in the CiviAI Discord Channel. I did not create this workflow; I used it as a template and modified it to suit my needs. My apologies for the videos. I will create a more PG-friendly set during the weekend. /Bow As some of you may know, I often use Pixel Dojo's reliable character creation tool to generate datasets for my Loras. However, each image costs one credit, and it usually takes around 25-30 images to create a dataset because the tool produces so many duplicate poses.  After watching one of Ostris's Lora training videos, I learned that he had created a WAN 2.2 I2v 360-degree Lora. This inspired me to consider a new approach: what if we could use a single image to create a rotating 360-degree video and then export each frame from that video using the workflow? This would come in handy for creating image datasets, and the images would remain consistent. And many of us could save money by generating datasets locally.

Files with SHA256: 15bfa9d4f0b155836e18ef3d0dccd28dae011144e78dd3563bd0d488846268b8