⚠️It is recommended to re-download the LoRA file for the JV1_ChillMix version if you downloaded it before December 24, 2024, or whenever I re-upload the file. This is because I occasionally update the dataset by adding or removing photos, as well as adjusting, improving, or refining the tags associated with the images.⚠️ Embarking on my first LoRA adventure! 🚀✨ Follow these recommended settings for optimal results:- DPM++ 2M Karras, Steps 25-50- Resolution: 512 x 768, 512x512- LoRA Weights: 0.6-1 (recommended 0.8)- Model: ChilloutMix or Majicmix, Realistic Vision may work, but be cautious as it may westernize faces- HiresFix: R-ESRGAN 4x+, 10-25 steps- Adetailer: face_<a target="_blank" rel="ugc" href="http://yolov8n.pt">yolov8n.pt</a>Note: I am not responsible for any NSFW or illegal content generated using this LoRA. Let's keep it creative and responsible! 🌐👩‍💻 #LoRA #AIart #ResponsibleAI 🔄 Important Settings Update for Automatic1111 Stable Diffusion WebUI! 🔄🚨 Issue Alert: If the LoRA name in your prompt doesn't match the LoRA filename, follow these steps for a quick fix:👉 In the 'Extra Network' settings, locate the option: "When adding to prompt, refer to LoRA by."👉 Change the setting to "filename."Explanation: A1111 introduced a new option with an unintended default value, causing the network picker to use the name from metadata (ss_output_name) rather than the filename in the prompt. This can make renaming LoRA files difficult. Switching to the correct value resolves this issue.

Niken

Trained with ChilloutMix checkpoint Altered some training images, fixing some flipped images

Files with SHA256: 537ea0c456defe7c059e3fa601d7a9c1e99116702fd0df4da30e97be2a12724a