Why is there no V4.0? Because I am not very satisfied with the 4.0 version, half of it was rejected by me...so I jumped directly to 5.0.
Some friends reported that when using the V3.0 version, it is not easy to make a full-body picture. After the weight is increased, it is difficult to control the movement of the characters. In addition, there are problems with the deterioration of the hands and the excessive impact on the background.
So this version is mainly used to fix these three problems.
V5.0 has improved in these three aspects, and it cannot be said to be a perfect solution, but at least it will not be so strenuous to adjust.
For example, no matter what base film you use, if you increase the weight of (full body), for example, to more than 1.7, there is a high probability that a full body image will be displayed.
When using this version, the weight can be relatively slightly larger. 0.65 is a critical point. Different weights have different effects.
In addition to the trigger word (cbpkv5), there are other words you can use, such as the following:
Character traits:
blue eyes,
breasts,
long hair,
lips,
small breasts,
multicolored hair,
green eyes,
mechanical ears,
nipples,
steampunk,
mecha musume,
mechanical parts,
robot joints,
single mechanical arm,
mechanical halo,
bodysuit,
mecha corset
Action and Posture:
standing,
legs apart,
hand on hip,
looking to the side,
cowboy shot,
ass focus,
Clothing & Accessories:
high heels,
headgear,
headset,
headphones,
mask,
Background and scene:
full body,
cable,
wire,
simple background,
reflective floor,
Other objects:
orb,
chain,
Note that these words do not have to be added, you can only use trigger words.
Because this is a stylized LORA, it will definitely have a certain impact on the background, not just the clothes. If you want to reduce the impact on other parts of the picture, you can open the layering settings and lower the weight of the MID layer (it is also possible to adjust it to 0. ), and at the same time increase the weight of this LORA, try different styles and compositions. (The V4.0 version can also be adjusted in this way. If the MID layer is adjusted to 0 in the V4.0 version, the whole body map can be produced very easily.)
Thank you very much for your love. This LORA upload has already had thousands of downloads in just a few days. I also have a ranking on station C. If you have any questions, please feedback and I will try my best to solve them.
About V5.0:
Why is there no V4.0? Because I am not very satisfied with the 4.0 version, half of it was rejected by me...so I jumped directly to 5.0.
Some friends reported that when using the V3.0 version, it is not easy to make a full-body picture. After the weight is increased, it is difficult to control the movement of the characters. In addition, there are problems with the deterioration of the hands and the excessive impact on the background.
So this version is mainly used to fix these three problems.
V5.0 has improved in these three aspects, and it cannot be said to be a perfect solution, but at least it will not be so strenuous to adjust.
For example, no matter what base film you use, if you increase the weight of (full body), for example, to more than 1.7, there is a high probability that a full body image will be displayed.
When using this version, the weight can be relatively slightly larger. 0.65 is a critical point. Different weights have different effects.
In addition to the trigger word (cbpkv5), there are other words you can use, such as the following:
Character traits:
blue eyes,
breasts,
long hair,
lips,
small breasts,
multicolored hair,
green eyes,
mechanical ears,
nipples,
steampunk,
mecha musume,
mechanical parts,
robot joints,
single mechanical arm,
mechanical halo,
bodysuit,
mecha corset
Action and Posture:
standing,
legs apart,
hand on hip,
looking to the side,
cowboy shot,
ass focus,
Clothing & Accessories:
high heels,
headgear,
headset,
headphones,
mask,
Background and scene:
full body,
cable,
wire,
simple background,
reflective floor,
Other objects:
orb,
chain,
Note that these words do not have to be added, you can only use trigger words.
Because this is a stylized LORA, it will definitely have a certain impact on the background, not just the clothes. If you want to reduce the impact on other parts of the picture, you can open the layering settings and lower the weight of the MID layer (it is also possible to adjust it to 0. ), and at the same time increase the weight of this LORA, try different styles and compositions. (The V4.0 version can also be adjusted in this way. If the MID layer is adjusted to 0 in the V4.0 version, the whole body map can be produced very easily.)
Thank you very much for your love. This LORA upload has already had thousands of downloads in just a few days. I also have a ranking on station C. If you have any questions, please feedback and I will try my best to solve them.
Thank you for your support.
关于V5.0:
为什么没有V4.0?因为4.0版本我自己不是很满意，做了一半被我否定掉了....所以直接跳到了5.0。
有朋友反馈，在使用V3.0版本时，不容易做出全身的图片，权重调大以后人物动作不好控制，另外还存在手部劣化，对背景影响过大的问题。
所以这个版本主要用来修复这三个问题。
V5.0在这三个方面均有所改善，不能说完美解决，但至少调整的时候不会那么费劲了。
比如，无论你用任何底膜，你如果把（full body）的权重调高，比如调整至1.7以上，极大概率会出全身图。
这个版本在使用时，权重可以相对稍微大一点。0.65是一个临界点。不同权重效果不一样。
除了触发词（cbpkv5）以外，您还可以使用其他的一些词语，举例如下：
人物形象特征：
blue eyes,
breasts,
long hair,
lips,
small breasts,
multicolored hair,
green eyes,
mechanical ears,
nipples,
steampunk,
mecha musume,
mechanical parts,
robot joints,
single mechanical arm,
mechanical halo,
bodysuit,
mecha corset
动作与姿势：
standing,
legs apart,
hand on hip,
looking to the side,
cowboy shot,
ass focus,
服饰配件：
high heels,
headgear,
headset,
headphones,
mask,
背景与场景：
full body,
cable,
wire,
simple background,
reflective floor,
其他对象：
orb,
chain,
注意，这些词不是说非要加进去，你只用触发词也是可以的。
因为这是一个风格化LORA，肯定会对背景有一定影响，而不是只影响衣服，如果希望减少画面其他部分的影响，可以打开分层设置，将MID层的权重调低（调成0也是可以的），同时调高本LORA的权重，尝试不同的风格与构图。（V4.0版本也可以这样调整，V4.0版本如果将MID层调成0，就可以非常容易的出全身图了）
非常感谢大家的厚爱，这个LORA上传短短几天已经有好几千的下载，我本人也在C站有了排名，有什么问题请反馈，我会尽量解决。
谢谢您的支持。
V5.0について：
V4.0 がないのはなぜですか? 4.0 バージョンにはあまり満足しておらず、半分は拒否されたため、直接 5.0 にジャンプしました。
一部の友人からは、V3.0 バージョンを使用すると、全身写真を作成するのが簡単ではない、重量が増加するとキャラクターの動きを制御するのが難しくなる、また、劣化の問題があるとの報告がありました。手の動きや背景への過度な影響。
したがって、このバージョンは主にこれら 3 つの問題を解決するために使用されます。
V5.0ではこの3つの点が改善されており、完璧な解決策とは言えませんが、少なくとも調整はそれほど大変ではなくなります。
例えば、どのようなベースフィルムを使用しても、（全身）のウェイトを例えば1.7以上にすると、高い確率で全身画像が表示されます。
このバージョンを使用すると、重量が比較的わずかに大きくなる可能性があります。
トリガー ワード (cbpkv5) に加えて、次のような他の単語も使用できます。
キャラクターの特性：
blue eyes,
breasts,
long hair,
lips,
small breasts,
multicolored hair,
green eyes,
mechanical ears,
nipples,
steampunk,
mecha musume,
mechanical parts,
robot joints,
single mechanical arm,
mechanical halo,
bodysuit,
mecha corset
動作と姿勢:
standing,
legs apart,
hand on hip,
looking to the side,
cowboy shot,
ass focus,
衣類とアクセサリー:
high heels,
headgear,
headset,
headphones,
mask,
背景とシーン:
full body,
cable,
wire,
simple background,
reflective floor,
その他のオブジェクト:
orb,
chain,
これらの単語は追加する必要はなく、トリガー ワードのみを使用できることに注意してください。
これは様式化された LORA であるため、服だけでなく背景にも確実に一定の影響を与えます。写真の他の部分への影響を軽減したい場合は、レイヤー設定を開いて MID のウェイトを下げることができます。レイヤー（0に調整することも可能です。）を追加し、同時にこのLORAのウェイトを増やし、さまざまなスタイルや構成を試してください。 (V4.0版でもこの調整は可能です。V4.0版ではMIDレイヤーを0に調整すると非常に簡単に全身マップが作成できます。)
愛していただき、誠にありがとうございます。この LORA アップロードは、わずか数日ですでに数千件のダウンロードがありました。ステーション C にもランキングがあります。ご質問がございましたら、フィードバックをお願いします。解決するために最善を尽くします。
ご協力ありがとうございます。