VTuber | Simfonira | RU
LORA
Pony
LORA
Pony
v1_Main

VTuber | Simfonira | RU

Azerlund
Creator
⭐ 0.0
⬇ 105 Downloads
👁 1 Views
🖼 18 Images

Trained Words

simfonira
source_reference
source_screen_good
model_type_fox
ears_fox
outfit_maid
maid
hair_tails
twintails
long hair
hair bow
eyes_purple
purple eyes
simfonira
source_reference
source_screen_good
model_type_cat
ears_cat
outfit_maid
maid
hair_long
very long hair
eyes_purple
purple eyes
simfonira
source_reference
source_screen_good
model_type_fox
ears_fox
outfit_pyjamas
pajamas
hair_tails
twintails
long hair
hair bow
eyes_blue
blue eyes
simfonira
source_reference
source_screen_good
model_type_cat
ears_cat
outfit_pyjamas
pajamas
hair_long
very long hair
eyes_blue
blue eyes
eyes_heart
heart-shaped pupils
score_9
score_8_up
score_7_up
source_anime
full body
standing
action pose
looking at viewer
dynamic angle

About this model

(●*∩_∩*●) LoRa for cutest VTuber Simfonira (づ。◕‿‿◕。)づ

Original Vtuber Links:

  • LORA was created with the express permission and personal request of the author.

    Like the model if you like it, this is important to me.

    ====

    Ready-made triggers from model details:

    • Copy and insert the whole. Everything will work at once.

    • 1 = HouseMaid-Fox - ideally

    • 2 = HouseMaid-Cat - ideally

    • 3 = Pajamas-Fox - Always canon clothes, but hair color and face 50/50. Tra random seeds or add general tags.

    • 4 = Pajamas-Cat - Always canon clothes, but hair color and face 50/50. Tra random seeds or add general tags.

    • 5 = Alternative love eyes, work.

    ====

    Original author: model: Custom tags below are required. Otherwise get just a similar anime girl.

    For good detail - near 18-22-25 steps. 14 is definitely little.

    The tag "full body" works normally.

    ====

    Just dump the list of all custom tags:

    Special general tags:

    • !!! source_reference - Original Highres Source Models. Priority use.

    • source_screen_good / source_screen_norm - Highres screenshots, Slightly low-quality.

    • source_art - High-quality 4 manual arts. Disputable.

    Special tags efficiency:

    • # model_type_fox / model_type_cat = Full work

    • # outfit_maid / outfit_pyjamas = Full work

    • tail_fox / tail_cat = Tail type in full body. In general, it works. Very little data.

    • ears_fox / ears_cat - Work

    • hair_long / hair_tails - Work, need additional tags.

    • eyes_purple / eyes_blue / eyes_heart - Work, need additional tags.

    ====

    Bonuses:

    • Есть шанс что 4 ручных арта немного отравили обучение.

    • In general, for all about 8 hours, under beer and on different days.

    • Time 5 in the morning but you need to finish... Finished at 8.

    ====

    Тех часть:

    • Dataset - 39 arts

    • 500 buzz and 50mins

    ====

    Not work / Useless:

    • jacket_orange / jacket_purple = External jacket. There was very little data. Does not work at all in any way.

    • headwear_maid / headwear_sleep = Caper or sleeping cap. On the cap was 1 art - total does not work.

    • model_2024 - general tag for the future if there are other models.

    ====

Related Models

Similar AI models you may like