Trained on YiffyMix v41 at 640px. Made with a newer method that produces images with similar colors and composition to the base model.
This negative embedding is trained on YiffyMix to produce higher quality images on that model. Use the embedding, "badyiffymix"(or badyiffymix35, badyiffymix41 for the newer versions), in the negative prompt to see a general improvement in quality, vibrancy, and style.
badyiffymix v1.0 - generates images with a flatter, shaded style
badyiffymix35 - generates images with more detail and realism while retaining badyiffymix style and Yiffymix image composition
badyiffymix41 - generates images with a less saturation and heavy contrast compared to the original model while generally creating a slight improvement in detail
I modeled this embedding after great negative embeddings like boring_e621 or easynegative. I also made it to see what I could do with chilon249's negative prompt:
low res, lowres, blurry, bad , letterbox, deformity, mutilated, malformed, amputee, watermark, signature, unusual , username, sketch, monochrome The newest version(badyiffymix41) doesn't use the above negative.
Trained on YiffyMix v41 at 640px. Made with a newer method that produces images with similar colors and composition to the base model.
1. 轉载模型僅供學習與交流分享,其版權及最終解释權归原作者。
2. 模型原作者如需認領模型,請通過官方渠道联系海藝AI工作人員進行認證。我們致力於保護每一位創作者的權益。 點擊去認領
