This is a LoRA for generating images of Lovelive character Takami Chika.
Making models takes a lot of time and hard work. If you like this LoRA, please click the like button!
Recommended weights:
Recommended weights: 0.5~0.7
(Note: If the weight is below 0.4, the LoRA may not work effectively.)
這是一個用於生成Lovelive角色高海·千歌的圖像的LoRA
製作模型需要花費很多的時間和精力。如果您喜歡這個LoRA，請幫忙按一個“讚”！
建議權重：
建議權重：0.5〜0.7
（注意：如果權重低於0.4，則LoRA可能無法生效。）
Description
FAQ
Comments (4)
I think I used too many pictures for training. It seems that training Lora doesn't require that many pictures. So now I'm considering using fewer pictures to speed up the process.
However, I'm not sure if using fewer pictures will negatively affect the final result or not.
As a fellow Love Live LoRa trainer, my datasets usually have between 25~30 handpicked images and that's that. What matters is the quality of the images and the tags, so go for it!
Thank you for your reply, your advice is very helpful!
Previously, I used at least 250 images for training, it seems that I really used too many images. Thank you again!
Yes, I also use around 30 images, that's enough. The most important thing I think is data quality.