v1.1 is dyLora. All the images are created with comfyui (without refiner).I copied and pasted prompts from the Civitai images for these images. Used this script from bdsqlsz- <a target="_blank" rel="ugc" href="https://civarchive.com/models/118694?modelVersionId=130494">https://civarchive.com/models/118694?modelVersionId=130494</a>

Samdoesarts SDXL

<ul><li>Increased training steps.</li><li>Increased training batch size (28)</li><li>Switched to dyLora with prodigy optimizerI trained for 40 epochs, but it became significantly overfit after 30 epochs.</li></ul>

Files with SHA256: 722cdebfe5d1512146cf05d8353e4ab6e3792584975cc9f560218b34ebad90bf