原创模型请勿盗用 ,请你标注出处

Please do not steal original models.Please mark the source

オリジナルモデルを盗用しないでください,出典をマークしてください

V3版版本已经发布请选择您合适的模型下载,v3版本是v1加v2版本的优化版本,,主要优化兼容性和无法风格化情况、以及权重的调节、推荐权重0.6-1.1之间，触发词：illustration,

The V3 version has been released, please choose your suitable model to download. The v3 version is an optimized version of the v1 plus v2 version. Mainly optimize compatibility and inability to stylize, as well as weight adjustment, recommend weight between 0.6-1.1, trigger word: illustration.

XL版本没有测试因为我的电脑配置不够运行 模型无法测试,请大家反馈给我情况

The XL version has not been tested because my computer configuration is not enough to run the model. Please give me feedback.

V2版本适合人物 v1版本适合风景

V2 version is suitable for characters v1 version is suitable for landscapes

这是我第六次炼丹了,询问大家喜欢,因为炼丹需要嗡嗡声,所以这个模型我会以兑换形式兑换，但是放心，是超低兑换金额,如果没有嗡嗡声我就无法进行工作,谢谢大家支持,

采样方法推荐：Euler a,迭代步数(Steps) 24-50之间

提示词

配套框架：

自然风格：插画、虫子、玫瑰、帽子、<lora:幻想童话插画风格_epoch_1:0.4>、宝石、亲密、动物、树、

【lora权重不能超过0.4】

否则超过0.4会出现的变异就会给你造成烦恼

人物风格：illustration,bug,flower,hat,<lora:幻想童话插画风格_epoch_1>,gem,border,1girl,wings,microdress,witch_hat,

【lora权重不能低于0.6】个人建议0.7 -1.4之间

否则会出现没有人物,

推荐图片尺寸

垂直方向：512*764

横向图片尺寸：764*512【不建议使用权限重超过1.4】因为训练是垂直方向尺寸会造成多人的情况可以添加反向关键词：多人

得到需要的效果,横向推荐权重是0.7-0.4之间

本模型建议不要使用修复功能,会破坏画面，

主要提示词是插图、亲密、

因为这是v1版本后续,如果有反馈，会更新版

This is my sixth time alchemy. I'd like to ask for your opinions. Because alchemy requires a humming sound, this model will be exchanged in the form of points, but don't worry, it's at an extremely low exchange rate. I can't work without the humming sound. Thank you for your support.

Sampling method recommendation: Euler a. Iteration steps (Steps) 24-50

Hints:

Compatible framework:

Natural style: illustration, insects, roses, hats, <lora: fairy tale illustration style_epoch_1> gemstones, intimacy, animals, trees,

[Lora weight cannot exceed 0.4]

Otherwise, the mutations that will arise due to exceeding 0.4 will cause you trouble.

Character style: illustration, bug, flower, hat, <lora: fairy tale illustration style_epoch_1>, gem, border, 1 girl, wings, microdress, witch hat,

[Lora weight cannot be less than 0.6] Personal recommendation 0.7 - 1.4

Otherwise, there will be no characters.

Recommended image size:

Vertical direction: 512*764

Horizontal image size: 764*512 (Not recommended to use Lora weight exceeding 1.4) Because training is in vertical direction size, it may cause the situation of many people. You can add the reverse keyword: many people to get the desired effect. Horizontal recommendation weight is 0.7 - 0.4

This model is not recommended to use the repair function. It will destroy the image,

The main hint words are illustration, intimacy,

Because this is the V1 version, subsequent updates will be made based on feedback.

これは私が6回目の练丹を试みたものです,皆さんの意见を侍いします,练丹には「バズ」が必要なので、このモデルは交换形式で提供しますが、安心してください、交换金额超低额です,バズがなければ、作业を进めることができません,皆さんのご支持を感谢します,

推奨サンプラング方法：Euler a,イテreーショnssuteppu数：24-50の间

ヒntワード
カードfureムワーク：
ナチュラルスタイル：插图，bug，花，帽子，lora：幻想童话eirasutosutairu_epoch_1：0.4，宝石，未知，动物，树，

キャラクターsutairu： 插图,bug,花,帽子,lora:幻想童话irasutosutairu_epoch_1,gem,border,1girl,wings,microdress,witch_hat,

推奨イメージサイズ
縦向き：512*764

横向きイメージサイズ：764*512（重みを1.4を超える使用は推奨しません）训练は縦向きサイズで行われ、复数人物のが表示されることがあります,逆のワード「多人」を追加することで、それを急できます,横向きは重みを0.7-0.4の间に推奨します,

このモデルは颜面修复机能の使用を推奨しません, 画面を破壊する可能性があります

描述:

待更新

待更新To be updated

因为由于本人的电脑无法运行XL模型 以上是云端训练。所以请大家及时反馈

Because my computer cannot run the XL model, the above is cloud training. So please give us your feedback in time.

训练词语: illustration