Sign In

Virtual Asian Female Character 20

24
218
254
7
Verified:
SafeTensor
Type
LoRA
Stats
140
1
Reviews
Published
Nov 11, 2024
Base Model
Flux.1 D
Training
Epochs: 1
Trigger Words
afg20
Hash
AutoV2
21035611D8
default creator card background decoration
XucroYuri's Avatar
XucroYuri
The FLUX.1 [dev] Model is licensed by Black Forest Labs. Inc. under the FLUX.1 [dev] Non-Commercial License. Copyright Black Forest Labs. Inc.
IN NO EVENT SHALL BLACK FOREST LABS, INC. BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH USE OF THIS MODEL.

Asian Girl Face:Training based on synthetic data.

Yes, perhaps you may think that this character is very similar to some real people in reality, but please believe me, it can only be a coincidence. The model training method I use is not a very special secret. At the beginning, I will have a training data set based on real people. There will be some pictures of real people classified according to face shapes. Then, through character consistency algorithms such as ipadpter, instantID, etc., a batch of new AI-generated characters that are equally beautiful, have some similarities but are actually significantly different are generated by mixing facial features of multiple characters. After accumulating a certain amount of materials, I will obtain some LoRA models of characters in the first stage. Then, in the second stage, in an AIGC environment similar to ComfyUI, I will randomly generate high-definition pictures with real photo texture and different occupations, clothing, expressions, postures, and environmental backgrounds by using nodes similar to SDXL Auto Prompter, etc., obtaining a larger and richer training set that is almost indistinguishable from real person photo materials. Well, after a new round of training, we have obtained a brand new character model. She is so beautiful, real, and has a rich and diverse life experience. But as you and I both know, she is a virtual character. You can use her in some research-oriented AIGC generation tasks, and you can also explore commercial uses. As long as you do not violate relevant laws and rules regarding content security and morality, there is no need to worry too much about unnecessary disputes caused by facial similarity.