Lora based on the model Niji Pony 3D Styles (TensorArt), I approached the 3D style through the Pony 🦄FaeTality in conjunction with the LoRA VividColor 2.5d [PonyXL] Style, bringing a cleaner style, with slightly strong lines and a reflection of light on the skin/hair of the art.

I mixed around 250 images to create LoRA in Civitai, making it more refined, especially in various poses, viewing angles, close-ups, shadow and light styles, male and female gender.

I decided to choose the Pony style because it was more versatile in terms of using tags, including a cartoon/realistic style. Among the other style options, I didn't reach a satisfactory conclusion, it was too cartoony or too realistic, so I decided to use Pony (in the future I intend to extend it to other SD 1.5 models, etc.).

Observation:

1) Use the tag hippo and/or hippopotamus

2) The tag slight chubby can sometimes make the character too obese, if you want to make a character not too fat or too thin, use the tag curvy figure

3) When using the tag close-up, especially on faces, it can make the character childish, if it is an adult character, in my tests that were successful, I used the tags ultra detailed, adult, mature

4) The tag tail had some errors in the hippo, such as a tail feline or reptile, so it is up to you to use the tail tag, in my tests, when I used the tail tag in the negative prompt, the model was doing the correct tail for a hippo

5) Use the tag 3D in the negative prompt so that the art does not look "plasticized" or in the "3D blender" style

Below is my default configuration in Civitai:

Model (base)

Pony 🦄FaeTality (v1.0)

LoRA (base)

VividColor 2.5d [PonyXL] Style (2.5d Realistic v1.0) - 0.4 ~ 0.8 (preferably 0.60)

Advanced

CFG Scale - 7.0 ~ 12.0 (preferably 7.0)

Sampler - Euler a ~ DPM++ 2M Karras (Any other type of sampler can have unexpected results)

Steps - 24 ~ 50 (preferably 24)

Clip Skip - 1 / 2 (preferably 1)

VAE - none

Prompt (Positive)

score_9, score_8_up, score_7_up, gray background, solo, solo_focus furry, anthro, female, hippo, nude, big breasts, wide hips, thick thighs, curvy figure, pose sexy, dynamic angle

Prompt (Negative)

zPDXL-neg, 3D, text, watermark, horns, tail

Here are the best results I've obtained using ComfyUI + DDetailer + Upscale:

Checkpoint

ckpt_name - ponyFaetality_v11.safetensors

CLIP Set Last Layer

stop_at_clip_layer - (-2)

LoraLoaderModelOnly

lora_name - 2.5dcolor-000009.safetensors

strength_model - 1.00

KSampler

steps - 25

cfg - 7.0

sampler_name - euler_ancestral / dpmpp_2m

scheduler - normal / karras

denoise - 1.00

Upscale Model

model_name - 4x_foolhardy_Remacri.pth

Latent Image

width - 1024

height - 1024

batch_size - 1

SAMLoader (Impact)

model_name - sam_vit_b_01ec64.pth

device_mode - auto

FaceDetailer (DDetailer)

guide_size - 512

guide_size_for - bbox

max_size - 1024

steps - 25

cfg - 7.0

sampler_name - euler_ancestral / dpmpp_2m

scheduler - normal / karras

denoise - 0.60

feather - 5

noise_mask - enabled

force_inpaint - enabled

bbox_threshold - 0.50

bbox_dilation - 20

bbox_crop_factor - 3.0

sam_detection_hint - center-1

sam_dilation - 0

sam_threshold - 0.93

sam_bbox_expansion - 0

sam_mask_hint_thresold - 0.70

sam_mask_hint_use_negative - False

drop_size - 10

Ultimate SD Upscale

upscale_by - 1.50

steps - 24

cfg - 7.0

sampler_name - euler_ancestral / dpmpp_2m

scheduler - normal / karras

denoise - 0.20

mode_type - Linear

tile_width - 1024

tile_height - 1024

mask_blur - 4

tile_padding - 32

seam_fix_mode - Band Pass

seam_fix_denoise - 1.00

seam_fix_width - 64

seam_fix_mask_blur - 8

seam_fix_padding - 16

force_uniform_tiles - true

Prompt (Positive)

score_9, gray background, solo, solo_focus furry, anthro, female, hippo, nude, big breasts, wide hips, thick thighs, curvy figure, pose sexy, dynamic angle

Prompt (Negative)

zPDXL-neg, 3D, text, watermark, horns, tail