HotDog Pony Mix XL版本V6 (ID: 820999)

A bunch of models i made based off of ponyV6, made usually with merges or custom loras. Since it's a merge based off of pony XL, the prompting rules are the same. Make shure to use

score_9, score_8_up, score_7_up, score_6_up...

V8: It's based on a newer muscle lora i made and the weird sausages mix, I mixed it into the hotdog models, and pushed it to a more schetcy style. I like the results. Like with all the hotog models, the character can come out a bit too buff, just add "Muscle" in the negative, or specify "Petite, skinny". You can use "Lips,nose" in the negative for a anime style, and be carefull about negatives, the model was made to use smaller prompts and negatives, so adding a essey into the negatives can mess things up. Also since the model is schetcy on purpose, if the face is far away it might get messed up, i reccomend working in higher resolutions for that.

V7: I really hated how V6 ended up oversaturated and janky, so i went back to V3.5 and changed it up with some stuff from V6. It's nice, better anatomy, solid for both anime and realism. Works better with shorter prompts and less negatives. Certain negatives like "graysacle, monochrome, broken limbs" can either overcook the model or give it a completley different look. Is what it is, you can play around with it.

V6: I went back to V4, merged it with some other models i made and pushed it to a more anime style. The model will fairly consistently have an anime stlye without "style_anime" or other tags, so it's closer to the V4 version, occasionally the anatomy might be jank, but the colors and style is overall nice.

V5: I made the model a bit more realistic, while trying to keep the nice colorus. It leans a bit more twords realism, but can easily be used for anime styles with keywords like "source_anime, lineart" or negatives like "lips, nose". It's a bit unstable, but flexible and fun to play around with.

V4: Redid the model from scratch with the idea to make it work better with anime styles. The model is much more flexible than the older versions, but it's a bit jank and inconsistent sometimes.

V3.5: Just pushed the model to a more "2.5D" look. I felt the V3 was too realistic. This version can still do a great job at realism as the previous version, but can do an anime style with more ease.

V3: The V3 version leans twords a semi realistic style, but can be used for anime as well. You can use keywords like "anime, anime screencap, source_anime,2d" to nudge the model to a anime style. Also, the model is great at drawing muscles, but sometimes you might get a character that's more buff than you would like, just add "muscular" to the negative, and you're good.

描述:

训练词语: