基于Anima Base 1.0训练 因为画师的有色图有一半是摸鱼图，所以衣服和头发会比较表现的比较平，背景越简单越明显，用自然语言生图和增加背景复杂度可以缓解 Trained on Anima Base 1.0 Since about half of the artist's colored illustrations are quick doodles, the details on clothing and hair may look relatively flat (this becomes more noticeable with simpler backgrounds). You can mitigate this by using natural language prompts and increasing background complexity.

Files with SHA256: 4c87a3990d00da6987abcbc3a00d1371289efe305a0e27ebac2d1f6fc5c9f6e3