If you have model customization or other cooperation or consultation needs, please contact WeChat: pomelonioi. Please simply indicate your intention

. Time is precious and don't chat casually. Thank you for understanding!

Hi everyone, I'm a Muggle! After more than 30 unsuccessful attempts, I got an SDXL model that has kept me excited for

a long time.

Based on HelloWorld_v2.0 (special thanks: LEOSAM is a rabbit), HotPot_sdXL has enhanced training on both [light and shadow] and [body].

Light and shadow materials enhance the strengths of SDXL and bring a very professional image texture. The shadows on the face, the backlight of the hair, and the outline of the characters all have a very advanced photographic lens feel;

The body material complements SDXL's shortcomings, enriching the character's posture and movements. The plump figure, slender legs, and the model's standard posture greatly enhance the expressive ability of a full-body picture; these two types of materials are just like a bite of mandarin duck hot pot, and on the other,

there is an elegant and beautiful Yang Chun Snow White is a grounded person on one side, and is perfect for portraying beautiful Chinese women with dark hair and dark eyes, hence the name Hotpot — Muggle Hot

Pot.

Here are tips for using the model: (1) Trigger words

mugglelight

can be used as a quality term to incorporate into your starting point;

(2) “Muggle Dream Light and Shadow” was uploaded to the personal website and [Muggle Spring Festival Time] Two new sets of wildcards, that is, I think, a good collection of reminders, and the supporting model works better; (3) sdxl is better at natural language recognition, and sd1.5 is no longer recommended to write prompts entirely using a single word.

You can refer to my example map or deck, and use the “one sentence to describe the subject of the screen+multiple words to add details”, which works better; (4) Models specialized for Asian women need to be enhanced in Europe, America, Africa, and men;

(5) Combination Turbo&LCM lora Rapid drawing can be achieved;

(6) Example parameters: sampler restart, resolution 832*1344 (HD repair 1.5 times, explosion memory can not be turned on or reduced), CFGScale 7-9 ( (Or use the Dynamic Thresholding -CFG Scale Fix plug-in) It is recommended to open adetailer for

full-body images or long-range

lenses. You can fill in the plug-in

: detailed face, detailed skin

(It doesn't have to be set this way; it's just my little personal habit) (6

) The VAE model is already included Set it, select None or

Auto. Never use SD1.5 VAE such as vae-ft-mse-840000, which will cause flower images!

If you think the model is easy to use, don't forget to give it to the model

Like+retweet, this is really important to me!

Also, for topics such as model training, drawing prompts, and the use of card drawing machines (automatic drawing), please join my exchange group 891417811