适用于Flux的动漫风格LORA，训练集是由AnythingXL生成的248张高质量图片，在flux dev上以batch size_3训练了10000步左右。由于这个LORA是我制作的第一个FLUX Lora，所以可能会有一些使用上的问题，如果你遇到任何问题，可以在评论区告诉我！ This is an anime-style LORA designed for Flux dev, the training set consists of 248 high-quality images generated by AnythingXL. And It was trained on Flux dev for about 10,000 steps with a batch size of 3. It's my first attempt at creating a FLUX Lora so there might be some issues when using it, if you encounter any problems, feel free to let me know in the comments section! I'd love to hear your feedback and improve it. 这是我的生成参数，可以作为参考： Here are my generation parameters and you can use them as a reference: <img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/7fb11e0b-f7b6-4412-a18f-b52fe08b64d9/width=525/7fb11e0b-f7b6-4412-a18f-b52fe08b64d9.jpeg" />———————————————————————————————— Lora的体积确实有点过大了，因为这是我的第一个FLUX Lora，所以我并不清楚将LORA size设定为多少才能比较好的还原画风。目前来看，将Lora的体积限制在100mb以内也可以很好地还原画风，下一个版本我会尝试更小体积的LORA。 I'll admit, the size of this Lora turned out to be a bit larger than expected. Since this is my first FLUX Lora, I wasn't quite sure what the ideal size setting would be to effectively capture the art style. From what I've seen so far, limiting the Lora size to under 100MB can still do a great job of reproducing the desired style. In the next version, I'm planning to experiment with an even smaller Lora size.

Files with SHA256: d63e77794506aab37ab8a98fb93f3205b6d4703ed8c38bdb0b0b55acd317bdf9