lu_XL soft cute little animal lora

0
0
0
0
RealisticAnimal PhotographyGirlStyle BoostObject EnhanceAnimal
Cập nhật gần đây: Đăng tải lần đầu:
Realistic,Animal Photography,Girl,LoRA,SDXLImage info
Realistic,Animal Photography,Girl,LoRA,SDXLImage info
Realistic,Animal Photography,Girl,LoRA,SDXLImage info
Realistic,Animal Photography,Girl,LoRA,SDXLImage info
Realistic,Animal Photography,Girl,LoRA,SDXLImage info
Realistic,Animal Photography,Girl,LoRA,SDXLImage info
Realistic,Animal Photography,Girl,LoRA,SDXLImage info
Realistic,Animal Photography,Girl,LoRA,SDXLImage info
Realistic,Animal Photography,Girl,LoRA,SDXLImage info
Realistic,Animal Photography,Girl,LoRA,SDXLImage info
Realistic,Animal Photography,Girl,LoRA,SDXLImage info
Realistic,Animal Photography,Girl,LoRA,SDXLImage info
Realistic,Animal Photography,Girl,LoRA,SDXLImage info
Realistic,Animal Photography,Girl,LoRA,SDXLImage info
Realistic,Animal Photography,Girl,LoRA,SDXLImage info
Realistic,Animal Photography,Girl,LoRA,SDXLImage info
Realistic,Animal Photography,Girl,LoRA,SDXLImage info

Lora Function: Enhance the cuteness and fluffiness of small animals.

【Material】Other non-SD AI models extracted, 18 images. Repeat 20, 6 epochs (basically default parameters: dim 64, alpha 32)

【Danlu】Liblib AI online, SDXL official model.

【Weights】0.6-0.8 (1.0 is also acceptable, but it may easily produce the effect of the material)

【Recommended Base Model】It is recommended to use base models in a true style, especially those like openxl that already come with cute animal data. Testing has shown that the effects of many base models are still good.

【Animal Types】Try to choose animals with fur (cat, dog, bear, fox, tiger, wolf, llama, leopard, manul...), or it will affect the fluffiness effect. Even though I tested a snake, the effect is as follows:

Please refer to the original picture example for parameters.


For other test effects, please refer to the xyz images below.

Model: openxlVersion30Movie_v30


Model: LEOSAM HelloWorld New World _ SDXL Realistic Big Model_v6.0


Model: Niji style XL_v1.1


Most materials are actually rabbits, but the generalization is also good. This experiment found that in SDXL learning, while learning the material to be 100% similar, the model can retain generalization by reducing the weight (0.6-0.8) + modifying the prompt words to produce a subject different from the material.

I originally thought it was a rabbit lora, but it turned out to be a soft cute animal lora.

If you want to see material rabbits, you can input: ribbonrabbit, weight 1.

ribbonrabbit, <lora:XL flufferabbit_rabbit_lora-06:1>, outdoors, in the wild,

Training tags for the labeled image are as follows:

Thảo luận

Phổ biến nhất
|
Mới nhất
Gửi
Đến sớm
Tải xuống
(0.00KB)
Chi tiết
Loại
Số lần tạo hình ảnh trực tuyến
0
Tải xuống
0
Tham Số Đề Xuất
Kết Hợp Đề XuấtCheckpoint
Không
Cân nặng được khuyến nghị.
0.8
CFG
5
VAE
Không

Bộ sưu tập hình ảnh

Phổ biến nhất
|
Mới nhất