한복 로라를 만들어주신 darong님과 Chals님께 먼저 감사드립니다.
best quality, cyberpunk, no_human, background
AND 1man, hanbok, <lora:hanbok_male:1>, <lora:Moxin_10:0.2>
Negative prompt: (out of frame, worst quality, low quality:1.4)
Steps: 20, Sampler: DPM++ 2M Karras, CFG scale: 7, Seed: 3306516825, Size: 912x512, Model hash: 30953ab0de, Model: meinamix_meinaV8, ControlNet-0 Enabled: True, ControlNet-0 Module: none, ControlNet-0 Model: control_openpose-fp16 [9ca67cc5], ControlNet-0 Weight: 1, ControlNet-0 Guidance Start: 0, ControlNet-0 Guidance End: 1, Latent Couple: "divisions=1:1,1.5:3 positions=0:0,0.5:1.5 weights=0.4,0.8 end at step=15"
best quality, cyberpunk, no_human, background
AND 1man, hanbok, shuimobysim, <lora:hanbok_male:1>, <lora:Moxin_10:0.6>
Negative prompt: (out of frame, worst quality, low quality:1.4)
Steps: 30, Sampler: DPM++ 2M Karras, CFG scale: 7, Seed: 529592995, Size: 912x512, Model hash: 1cb9f9c58a, Model: anythingworld_v15, ControlNet-0 Enabled: True, ControlNet-0 Module: none, ControlNet-0 Model: control_openpose-fp16 [9ca67cc5], ControlNet-0 Weight: 1, ControlNet-0 Guidance Start: 0, ControlNet-0 Guidance End: 1, Latent Couple: "divisions=1:1,1.5:3 positions=0:0,0.5:1.5 weights=0.4,0.8 end at step=20"
best quality, cyberpunk, no_human, background
AND 1girl, hanbok, shuimobysim, <lora:hanbok-000009:1> <lora:Moxin_10:0.8>
Negative prompt: (worst quality, low quality:1.4)
Steps: 30, Sampler: DPM++ 2M Karras, CFG scale: 7, Seed: 3702575565, Size: 912x512, Model hash: 1cb9f9c58a, Model: anythingworld_v15, ControlNet Enabled: True, ControlNet Module: none, ControlNet Model: control_openpose-fp16 [9ca67cc5], ControlNet Weight: 1, ControlNet Guidance Strength: 1, Latent Couple: "divisions=1:1,1.5:3 positions=0:0,0.5:1.5 weights=0.4,0.8 end at step=20"
best quality, cyberpunk, no_human, background
AND 1girl, hanbok, shuimobysim, <lora:hanbok-000009:1> <lora:Moxin_10:0.6>
Negative prompt: (worst quality, low quality:1.4)
Steps: 30, Sampler: DPM++ 2M Karras, CFG scale: 7, Seed: 3304662813, Size: 912x512, Model hash: 30953ab0de, Model: meinamix_meinaV8, ControlNet-0 Enabled: True, ControlNet-0 Module: none, ControlNet-0 Model: control_openpose-fp16 [9ca67cc5], ControlNet-0 Weight: 1, ControlNet-0 Guidance Start: 0, ControlNet-0 Guidance End: 1, Latent Couple: "divisions=1:1,1.5:3 positions=0:0,0.5:1.5 weights=0.4,0.8 end at step=20"
구도는 컨트롤넷의 openpose로 잡았습니다. 투샷을 활용해서 배경과 캐릭터의 느낌을 나눠보려고 했습니다만 잘 적용된건지 모르겠네요.
추가로 전에 투샷 활용방법 글에서 아래나 오른쪽의 절반보다 큰 영역을 지정하는 법이 없어보인다고 했는데, 잘못된 설명이었네요.
참고 : https://www.clien.net/service/board/cm_aigurim/17933703CLIEN
아래처럼 Position에 소숫점을 사용하면 지정이 가능했습니다.
아래 이미지들은 덤입니다. 수묵화 느낌이 어울릴 것 같아서 MoXin이라는 중국 수묵화 로라를 섞어 테스트 해본 것입니다.
https://civitai.com/models/12597/moxin