No Trigger.
About 0.8-1.0 Change the value according to the distance of the subject.
I tried to reduce the file size, but I couldn't do it because the drawing became vague.
It often works with other 2D base models as well.
If you want to reduce the influence of the drawing style, please use [LoRA Block Weight].
After installing LoRA Block Weight, please specify the Lora name as → <lora:**********:1:lbw=0,0,0,0,0,0,0,0,1,1,1,1,0,0,0,0,0>
Please change the place of 1, 1, 1, 1, and correct it.
We have newly released 3way v2.
Adjusted to make the nostrils bigger.For front only. The accuracy of reproduction decreases as the subject moves away.
Mainly works on the upper body and close range.
V1 is front only.
The V2 works even in slightly oblique conditions. I am using different data from V1.
3way v1.
V4 has not been released from the beginning.
V5 was retrained by integrating the images of V1, V2, and V4.
Nose-Tape V1.0 is intended to be used to attach tape to the nose and pull it up. little version up.
V6. The learning size was set to 768 x 768.Mainly works on the upper body and close range.
Newer versions aren't necessarily better. Each has different characteristics.
Use Base model: Counterfeit-V3.0_fp32
Please use them according to the situation.
I don't know if it can be used well because it was made by an AI beginner.
Whole-body image generation requires long prompts. It will generate a lot of weird images until it works.
Please refer to the prompt used for the V5 image and make it.
No trigger word required.
I would appreciate it if you could upload a good image.
Commercial use is basically prohibited. We will not be held responsible even if any trouble occurs after the product is sold. Please give priority to settings that comply with the rules.
描述:
About 0.8
V1のデータを修正して学習しなおしました。
主に上半身でしか機能しません。顔は正面のみ適用されます。
単体で使うと変な色合いになる場合が多いです。口が閉じません。
他にも問題点は残っています。
训练词语:
名称: NOSEHOOKv2.6.safetensors
大小 (KB): 295033
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success
