Type | |
Stats | 1,141 |
Reviews | (92) |
Published | Nov 20, 2024 |
Base Model | |
Hash | AutoV2 F09517269F |
❤If you like this model, just give me some emojis, thanks❤
You can use the generated images for commercial purposes, without my permission.The reason I published this model is to share it with others. It would be best if you would like to keep the original name and address of the model when using it.
Note that I will not bear any legal responsibility for the pictures you generate. Please use them as permitted by local law.
=======================================================
Brief description (If you don't want to read the all content , it is recommended to read this short paragraph,But I suggest you read it all, there are important instructions)
If you can't see the image, it's because it's been reviewed.
If the face in the full-body image is distorted, please use FaceDetailer or ADetailer to fix it
KSampler Settings:
steps:25-35
cfg:4.5-7
sampler:Dpmpp_2m karras or Euler_ancestral karras
resolution:
I usually use a rectangular resolution to draw people(Non-NSFW image),For example, 896x1344.
prompt words:
For best results, use the prompt words I suggest
Prompt:
score_9,score_8_up,score_7_up,
1girl,perfect_face,exquisite_face,perfect_ratio_proportioned_face,
The following is the prompt word you need (short_hair,narrow_waist, sitting, reading, library, etc.)
NFSW should consider put the action prompts first
Negative prompt:
score_4,score_5,score_6,messy_drawing,amateur_drawing,bad_drawing,ugly_eyes,worst_quality,low_quality,normal_quality,low_resolution,child,toddler,infant,cherub,college_age,middle_age,grandmother,mother,milf,teens,teeth
Image quality enhancement prompt words (optional):
high_definition_with_hyper-realistic_details,portrait_taken_with_a_50mm_lens,intricate_details,realistic_textures,soft_natural_lighting,highlighting_features,vivid_colors_and_lifelike_shadows,crisp_focus_on_the_subject_with_soft_bokeh,shallow_depth_of_field,softly_blurred_background,dramatic_contrasts,cool_lighting,shadow_play,chiaroscuro_effect,warm_tones,cool_shadows,diffused_light,colored_gels,cinematic_flare,dynamic_range,soft_fills,atmospheric_haze,visual_narrative
Example:
score_9,score_8_up,score_7_up,face_focus,looking_at_viewer,
1girl,perfect_face,exquisite_face,perfect_ratio_proportioned_face,sitting,head_tilt,pose,long_hair,(cappuccino \(drink\):0.4), (table,small_cakes:0.6),indoors,(street:0.2),large_breasts,(mature_female:0.5),strapless_dress,
high_definition_with_hyper-realistic_details,portrait_taken_with_a_50mm_lens,intricate_details,realistic_textures,soft_natural_lighting,highlighting_features,vivid_colors_and_lifelike_shadows,crisp_focus_on_the_subject_with_soft_bokeh,shallow_depth_of_field,softly_blurred_background,dramatic_contrasts,cool_lighting,shadow_play,chiaroscuro_effect,warm_tones,cool_shadows,diffused_light,colored_gels,cinematic_flare,dynamic_range,soft_fills,atmospheric_haze,visual_narrative
Prompts Tips:
The order of prompt words is very important. As we all know, PONY is not good at understanding natural sentences and needs to write Danbooru words, so you need to pay attention to the order of your prompt words. Put the important prompt words of the picture in front. If it is a NFSW picture, put the action after 1girl.
The order of my picture prompt words is generally as follows:
1girl---(Facial features of the character, including fixed words: perfect_face, exquisite_face, perfect_ratio_proportioned_face. Physical features of the character, including large_breasts, narrow_waist, etc.) - (character's posture) --- (character's clothing, including accessories, etc.) --- (background of the picture, weather, time, etc.).
Of course, there are exceptions. For example, if I need to draw a picture of a woman with her hair combed back, hair_slicked_back cannot achieve the desired effect. In this case, you need to add the words forehead and earrings. In this case, you need to write hair_slicked_back, forehead and earrings after 1girl:
1girl, hair_slicked_back, forehead, earrings
Upscale Sittings:
scale by:1.2
denoise:0.55
steps:30
cfg:The same CFG that you used to generate the image
Sampler:The same Sampler you used to generate the image
=======================================================
The original version of this model was merged from the following models:
DucHaiten-Pony :https://civitai.com/models/376450/duchaiten-pony-xl-no-score
real pony:https://civitai.com/models/365041/real-pony
Please support the original author of the model.
This model is merged using Comfyui ModelMergeSDXLDetailedTransformers, each parameter is adjusted in detail, and own training content is added.
Realistic portrait photography style,NFSW Support.
Best used with Comfyui.
Just use the basic workflow. If you need to use ControlNet or FaceDetailer, I also provide an integrated workflow,Download Comfyui WorkFlow here:
https://civitai.com/articles/5579
============================================================