Type | |
Stats | 1,146 |
Reviews | (276) |
Published | Apr 2, 2023 |
Base Model | |
Training | Steps: 31,680 Epochs: 12 |
Trigger Words | tac001_style blurry chromatic_aberration depth_of_field floating_hair liquid see-through |
Hash | AutoV2 985A2CF23C |
What is Top Artist Challenge?
i always heared “AIGC have not soul.”
so,i just want to try this style lora series .
i will use many ways to control best effect on AIGC.
you can find tac000 in here
For certain reasons, I have decided to hide the name of the artist and use a code name instead.
Who is 001?↓
i dont know,maybe ....
DIFF form LoRA Version:
advance version LoRA,don't need Improved quality prompt such as 'master piece,best quality',
it can show train data quality.
it can change to other model such as 3D real and works well.
BUT
LoHA cant load with Additional Networks
you must use
<lora:[TAC]tac001_style_LoHA_V1:0.8>
to trigger model
you can see more info in LyCORIS
Trigger words:
tac001_style
blurry
chromatic_aberration
depth_of_field
floating_hair
liquid
see-through
001's painting style has four 2 characteristics, the most obvious are the Dynamic and Lens feeling usage.
you can try to get Dynamic with these trigger words:
floating_hair
liquid
you can try to get Lens feeling with these trigger words:
chromatic_aberration
blurry
depth_of_field
I made some adjustments specifically for transparent scenes, and the next version will differentiate between close-up facial training and full-body training.
Extra
if you want to get more color use,you can try
vivid_color
neno_color
I did not specifically train for these two prompts, but I still achieved good results.
and use monochrome as negative prompt.
i upload train data caption txt,you can use it with wildcard for test random effects.
just put it to extension/wildcard dir and use __tac001__ to try