v1.1: Fixed overtraining (i hope). It was pretty obvious looking back at the images but I don't have a lot of time to train/test. I will rely mostly on you guys to experiment and let me know in the comments. This model I had to fully retrain but on all future models I will be saving all model epochs until I run out of storage space
Trained in local Kohya install. 4090. Trained on DreamShaper XL1.0 Alpha2. I followed SECourses SDXL LoRA Guide.
Most images were on DreamShaper XL A2 in A1111/ComfyUI. Hi-res fix with R-ESRGAN (1.25) and 0.2-0.4 denoising strength. Upscaled using "4x_NickelbackFS_72000_G" or "4x_NMKD-Siax_200k"
Help me out by liking and following me to see my new models. Submit some of your outputs and post your SUGGESTIONS in the comments below. I am pacing around 6-10 models weekly depending on my actual job workload. If you liked my models and want to support the work I do you can buy me a pizza!
List of upcoming LoRAs I intend to do as well as what I am currently working on.
Description
FAQ
Comments (2)
Looking good. I've tried 36-120 images with varying results, usually 10k steps. What's your preferred number of images and steps? Looking forward to the rest on your list!
For most of my LoRAs I use around 15-20 images in varying outfits, expressions and backgrounds. Usually between 6k-10k steps. 😁
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.