Instructions to use althjs/sig_t_elderly_female_eastern with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use althjs/sig_t_elderly_female_eastern with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("black-forest-labs/FLUX.2-klein-4B", dtype=torch.bfloat16, device_map="cuda") pipe.load_lora_weights("althjs/sig_t_elderly_female_eastern") prompt = "hejin_sig_t, close-up portrait of an elderly Korean woman, soft pastel sweater, calm smile, Studio Ghibli watercolor" image = pipe(prompt).images[0] - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- Draw Things
- DiffusionBee
sig_t_elderly_female_eastern
60๋+ ๋์ ์ฌ์ฑ ยท ํ๊ตญ ํ ๋จธ๋ / ์ฌ๊ทน ๋ ธ๋ถ์ธ
ํ๊ตญ YouTube ์ฝํ ์ธ ์๋ํ ํ์ดํ๋ผ์ธ hejin ์ ์๊ทธ๋์ฒ ์บ๋ฆญํฐ ๋ผ์ด๋ธ๋ฌ๋ฆฌ LoRA.
ํต์ฌ ์ ๋ณด
| ํญ๋ชฉ | ๊ฐ |
|---|---|
| Trigger word | hejin_sig_t |
| ๊ถ์ฅ strength | 0.8 |
| Base model | Flux 2 Klein 4B |
| ๋ผ์ด์ผ์ค | apache-2.0 |
| ํ์ผ๋ช | sig_t_elderly_female_eastern_klein4b_v1.safetensors |
์บ๋ฆญํฐ ์ ๋ณด
์ธ๋ชจ ๋ฒ ์ด์ค: a woman in her late 60s to 70s East Asian (Korean) with warm wise gentle presence, white or silver-grey hair softly tied in a low bun, deep lines around kind warm eyes filled with wisdom, refined face with mature soft features, slightly hunched gentle build, often a soft warm smile. ๊ถ์ฅ ์์ prior: hanbok ๋ ธ๋ถ์ธ ์์ (์ด๋์ด ์์กฐ์ ๊ธด ์น๋ง์ ๊ณ ๋ฆฌ) / ํ๋ณต (hanbok Korean traditional dress) elderly style / ํ๊ตญ ํ ๋จธ๋ ์ท (modern). ๊ถ์ฅ ๋ถ์๊ธฐ: ํ๊ตญ ์ฌ๊ทน์ ๋์๋๋น/๋ ธ๋ถ์ธ, ํ๊ตญ ํ ๋จธ๋, ๋ง์ ์ด๋ฅธ, ์ ์์ด. ํ์ฉ ์ํ ์: ํ๊ตญ ์ฌ๊ทน์ ๋์๋๋น, ๋ํ ์ ํ ๋จธ๋, ๋ง์ ์ด๋ฅธ. dataset ์์ฑ prompt ๋ฒ ์ด์ค: 'an elderly East Asian woman with silver hair softly tied, deeply lined warm wise eyes, refined gentle face, soft kind expression, painterly Studio Ghibli watercolor, soft warm interior light.'
ํ์ต ์ ๋ณด
- ๋ฐ์ดํฐ์ : 20 ์ปท (1024ร1024) โ ์์ / ํ์ / ๋ฐฐ๊ฒฝ ๋ค์ํ
- Caption ํ์:
hejin_sig_t, <framing> of <identity>, <์์ 1~2 ๋จ์ด>, <ํ์ 1~2 ๋จ์ด>(60~90 chars sweet spot) - Caption ์์น: face ๋ฌ์ฌ (๋์ยท๋จธ๋ฆฌ ๊ธธ์ดยทํผ๋ถ ๋ฑ) ๋ caption ์ ๋ฐ์ง ์๊ณ trigger word ๋ง์ผ๋ก face ํธ์ถ โ face ํ์ต ์ ํธ dominant
- Base model: Flux 2 Klein 4B
- ํ์ต ๋๊ตฌ: ai-toolkit (Ostris) โ rank 24 / alpha 16 / lr 1e-4 / steps 1500 / flowmatch scheduler
์ฌ์ฉ ์์ โ WaveSpeed Cloud API
curl -X POST https://api.wavespeed.ai/api/v3/wavespeed-ai/flux-2-klein-4b/text-to-image-lora \
-H "Authorization: Bearer $WAVESPEED_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"prompt": "hejin_sig_t, close-up portrait of an elderly Korean woman, soft pastel sweater, calm smile, Studio Ghibli watercolor",
"size": "1024*1024",
"seed": 12345,
"loras": [{"path": "https://huggingface.co/althjs/sig_t_elderly_female_eastern/resolve/main/sig_t_elderly_female_eastern_klein4b_v1.safetensors", "scale": 0.8}]
}'
๋ผ์ด์ผ์ค
์ด LoRA ๋ base model Flux 2 Klein 4B ์ ๋ผ์ด์ผ์ค (apache-2.0) ๋ฅผ ๋ฐ๋ฅธ๋ค.
Apache 2.0 โ ์์ ์ฌ์ฉ / ์์ / ์ฌ๋ฐฐํฌ ์์ . royalty-free.
ํ์ต ๋ฐ์ดํฐ์ ๋ฏธ๋ฆฌ๋ณด๊ธฐ
ํ์ต์ ์ฌ์ฉ๋ dataset ์ผ๋ถ (์ ์ฒด 20์ปท์ dataset/ ํด๋ ์ฐธ์กฐ):
๊ฐ ์ปท์ caption ํ์ผ (dataset/NN.txt) ์ ํ์ต ์ ์ฌ์ฉ๋ ์งง์ ํ์ (60~90 chars) ์ผ๋ก trigger + framing + identity + ์์ + ํ์ ๋ง ํฌํจ.
- Downloads last month
- 9
Model tree for althjs/sig_t_elderly_female_eastern
Base model
black-forest-labs/FLUX.2-klein-4B


