Type | |
Stats | 264 0 |
Reviews | (47) |
Published | Mar 24, 2023 |
Base Model | |
Trigger Words | sirius 1girl solo |
Hash | AutoV2 8C181B150D |
这是我训练的第二个Lora模型。这个角色是Re:creator的席利乌斯,仅仅出场一集她后就被阿尔泰尔占据了身体。
这个Lora模型是基于final-prune.ckpt(【旧短哈希1d4a34af】新短哈希22fa233c2d),使用了768x768的数据集进行训练。
通常情况下不需要Trigger Words就能触发该Lora,但是为了更好的效果,还是推荐加上sirius,1girl,solo这些Trigger Words。
推荐使用0.6-0.8的权重,除了提到的触发词以外,其他的人物特征是不必要提到的,但如果你想要更好的效果,也可以添加,因为一些细节(比如人物的袖章,或者护手可能会在生成时丢失)。
如果人物无法生成你想要的动作,请增加动作的权重或者降低Lora的权重。
如果生成的人物图像面部不佳,推荐您勾选高清修复。
这个模型可以生成nsfw的图片。
测试了kotosmix以及AbyssOrangeMix2,AbyssOrangeMix3,以及anyhentai和Grapefruit,其中AbyssOrangeMix的模型生成的人物图像面部,可能会出现一些问题。
如果可以的话,请在评论区留言以告知我这个Lora在其他的模型上是否好用,谢谢。
英语是用Google翻译的,如果有造成误解之处非常抱歉。
This is the second LORA model I trained. This role is Re: Creator's Sirius, after just one episode, she was occupied by Altair.
This LoRa model is based on Final-Prone.ckpt ([Old Short Hash 1D4A34AF] New Short Hash 22FA233C2D), which uses a dataset of 768x768 for training.
Usually, the TRIGGGER WORDS is required to trigger the LoRa, but for better results, it is recommended to add Sirius, 1Girl, SOLO these Trigger Words.
It is recommended to use the weight of 0.6-0.8. Except for the triggers mentioned, other character characteristics are not necessary to mention, but if you want a better effect, you can also add it, because some details (such as the character's sleeve, Or the handle may be lost during generation).
If the character cannot generate the action you want, please increase the weight of the action or reduce the weight of LORA.
If the character's image is not good, it is recommended to check the high -definition repair.
This model can generate NSFW pictures.
Test Kotosmix and ABYSSORANGEMIX2, ABYSORANGEMIX3, and Anyhentai and Grapefruit, where the character image of the character generated by ABYSSARANGEMIX may have some problems.
If possible, please leave a message in the comment area to let me know if this Lora works well on other models, thank you.
English is translated with Google, which is very sorry for misunderstanding.