Type | |
Stats | 455 228 |
Reviews | (85) |
Published | Sep 4, 2023 |
Base Model | |
Trigger Words | mauxirA1CB,1girl, solo, standing, naughty face, smile, look at viewer, headgear, yellow eyes, sparkling eyes, mole under eye, dress, gloves, long hair, low twintails, cape, belt, long sleeves, high heel boots, socks, thigh strap, detailed face |
Hash | AutoV2 30093A2FB0 |
权重：0.6-0.8，推荐0.65-0.72，权重越高，3D感越强
推荐使用：高分辨率修复 (Hires. fix)、IMAGE-TO-INAMGE / controlnet-tile 放大重绘
【推荐使用模型】：V2- OnlyAnime |《唯》· 炫彩动漫 https://civitai.com/models/105955/onlyanime-or
【负面embaddings】:EasyNegative、bad-hands-5
CFG:7
vae: kl-f8-anime2
本次不推荐使用UnrealisticDream，因为使用后会导致眼睛的出现错误的概率上升。
——————————————————————————————————————————————————————
【Trigger Words】
1.正面提示词：mauxirA1CB, 1girl, headgear, yellow eyes, mole under eye, dress, gloves, long hair, low twintails, cape, belt, long sleeves
正面可选提示词：sparkling eyes, naughty face, smile, high heel boots, socks, thigh strap, detailed face
2.负面提示词：3D,3D rendering, weapon, holding, holding weapon, gun, holding gun,
负面可选提示词：poor light
3.负面embaddings: EasyNegative,bad-hands-5
【一个参考预设，您可以在此之上进行修改】
1.正面提示词：mauxirA1CB,1girl, solo, standing, naughty face, smile, look at viewer, headgear, yellow eyes, sparkling eyes, mole under eye, dress, gloves, long hair, low twintails, cape, belt, long sleeves, arms behind back, (cowboy shot:1.1), sunlight, outdoors, masterpiece, best quality, (colorful),(finely detailed beautiful eyes and detailed face),cinematic lighting, extremely detailed CG unity 8k wallpaper, <lora:mauxirA1CB-v1:0.7>
2.负面提示词：EasyNegative, 3D, 3D rendering, weapon, holding weapon, gun, poor light, bad-hands-5
3.其它参数：DPM++SED Karras, 40 steps ,512x1024， 启用Hires.fix, R-ESRGAN 4X+Anime6B, 放大1.25倍，重绘幅度0.3-0.4
——————————————————————————————————————————————————————
【目前存在的问题与建议】
1.（1）在绘制远景镜头的图片时（full body），人物的眼睛与面部精度有概率出现瑕疵，但可以用以下方式【3】修复。
（2）人物头饰、胸前配饰、披风和腿部的服饰有概率出现错误，但不会影响整体人物观感。
（3）可以尝试使用 sparkling eyes, naughty face, smile 等提示词增强人物观感。
（4）在512x768尺寸cowboy_shot, upper_body的图片中，人物的头饰可能会变短。
2.在绘制远景或全身画面时，请适当增大出图尺寸与步数，这有助于提升人物面部精度。
3.如果以上方法【2】的效果不理想，请直接将生成的图片发送至image-to-image进行放大重绘【效果很好】，方法如下：
（1）尽量使用与原图一致的提示词、角色lora、seed
（2）启用controlnet--tile, 启用 放大脚本 Ultimate SD upscale
（3）一个参考设置参数：Steps:30-40, 重绘幅度0.25-0.4，放大算法：4x-UltraSharp 或 R-ESGAN 4x+anime6B, 放大倍数1.25--
4.在创作NSFW内容时，请根据实际情况降低权重至0.5-0.65，同时使用更多的NSFW相关提示词，增加NSFW相关提示词的权重，并将dress, gloves,cape, belt, long sleeves, high heel boots, socks, thigh strap等衣服类Trigger Words 加入负面提示词。
5.当您感觉图片细节不足、模糊时，请适当增加迭代步数，或调整采样器及vae,也可以尝试【3】的方法进行图生图重绘。
6.强烈推荐使用负面embaddings:EasyNegative
7.可能存在的模型泛用性问题
__________________________________________________________________________________________________________________________________________________________________
Weight: 0.6-0.8, recommended 0.65-0.72, weight, the stronger the 3D sense
【Recommended use】: Hires.fix, IMAGE-TO-INAMGE / controlnet-tile zoom-in redraw
【Recommended model】: V2- OnlyAnime |《唯》· 炫彩动漫 https://civitai.com/models/105955/onlyanime-or
【Negative emaddings】EasyNegative、bad-hands-5
CFG:7
vae: kl-f8-anime2
It is not recommended to use UnrealisticDream this time because using it will increase the probability of eye errors.
——————————————————————————————————————————————————————————————————————
【Trigger Words】
1. Positive prompt words: mauxirA1CB, 1girl, headgear, yellow eyes, mole under eye, dress, gloves, long hair, low twintails, cape, belt, long sleeves
Optional prompt words on the front: sparkling eyes, naughty face, smile, high heel boots, socks, thigh strap, detailed face
2. Negative prompt words: 3D, 3D rendering, weapon, holding, holding weapon, gun, holding gun,
Negative optional prompt word: poor light
3. Negative embaddings: EasyNegative,bad-hands-5
【A reference preset from which you can make modifications】
1. Positive prompt words: mauxirA1CB,1girl, solo, standing, naughty face, smile, look at viewer, headgear, yellow eyes, sparkling eyes, mole under eye, dress, gloves, long hair, low twintails, cape, belt, long sleeves, arms behind back, (cowboy shot:1.1), sunlight, outdoors, masterpiece, best quality, (colorful),(finely detailed beautiful eyes and detailed face),cinematic lighting, extremely detailed CG unity 8k wallpaper, <lora:mauxirA1CB-v1:0.7>
2. Negative prompt words: EasyNegative, 3D, 3D rendering, weapon, holding weapon, gun, poor light, bad-hands-5
3. Other parameters: DPM++SED Karras, 40 steps, 512x1024, enable Hires.fix, R-ESRGAN 4X+Anime6B, amplify 1.25 times, redraw range 0.3-0.4
————————————————————————————————————————————————————————————————————————————————————————————————————————————————
[current problems and suggestions]
1. (1) When drawing long-shot pictures (full body), there is a probability that the character’s eyes and facial accuracy will have flaws, but they can be repaired in the following ways [3].
(2) There may be errors in the character's headgear, chest accessories, cloak and leg clothing, but it will not affect the overall look and feel of the character.
(3) You can try to use sparkling eyes, naughty face, smile and other prompt words to enhance the look and feel of the character.
(4) In the 512x768 size cowboy_shot, upper_body picture, the character's headgear may become shorter.
2. When drawing distant views or full-body scenes, please increase the drawing size and number of steps appropriately, which will help improve the facial accuracy of the characters.
3. If the effect of the above method [2] is not satisfactory, please directly send the generated image to image-to-image for enlargement and redrawing [the effect is very good], the method is as follows:
(1) Try to use prompt words, characters lora, and seed that are consistent with the original picture.
(2) Enable controlnet--tile, enable the amplification script Ultimate SD upscale
(3) A reference setting parameter: Steps:30-40, redrawing range 0.25-0.4, magnification algorithm: 4x-UltraSharp or R-ESGAN 4x+anime6B, magnification factor 1.25--
4. When creating NSFW content, please reduce the weight to 0.5-0.65 according to the actual situation, and at the same time use more NSFW-related prompt words, increase the weight of NSFW-related prompt words, and use dress, gloves,cape, belt, long sleeves, Add negative prompt words to Trigger Words for clothes such as high heel boots, socks, thigh strap, etc.
5. When you feel that the details of the picture are insufficient or blurry, please increase the number of iteration steps appropriately, or adjust the sampler and vae, or try the method [3] to redraw the picture.
6. It is strongly recommended to use negative emaddings:EasyNegative
7. Possible model versatility problems