LoRA κΈ°λ° Stable Diffusion v1.5 μλμ°¨ μ΄λ―Έμ§ νμΈνλ λͺ¨λΈ π
μ΄ λͺ¨λΈμ sd-legacy/stable-diffusion-v1-5
λ₯Ό κΈ°λ°μΌλ‘,
νλμλμ°¨ μ΄λ―Έμ§μ ν΄λΉ λμμΈμ μ€λͺ
ν ν
μ€νΈλ‘ LoRA λ°©μμ νμΈνλμ μ§νν λͺ¨λΈμ
λλ€.
μ΄ 58κ°μ μ΄λ―Έμ§-ν
μ€νΈ μμ μ¬μ©ν΄ νμ΅μ μ§ννμΌλ©°, train_text_to_image_lora.py
μ€ν¬λ¦½νΈλ₯Ό νμ©νμ¬ μλμ κ°μ μ€μ μΌλ‘ νμ΅νμ΅λλ€.
π§ νμ΅ μ€μ
- ν΄μλ: 512 x 512
- λ°°μΉ ν¬κΈ°: 1 (gradient accumulation steps = 4)
- μ΄ νμ΅ μ€ν μ: 1000 steps
- 체ν¬ν¬μΈνΈ μ μ₯ μ£ΌκΈ°: 100 stepλ§λ€ μ μ₯
- νΌν© μ λ°λ: fp16
- μλ κ°: 42
π‘ μ¬μ© λ°©λ²
from diffusers import StableDiffusionPipeline
pipe = StableDiffusionPipeline.from_pretrained(
"sd-legacy/stable-diffusion-v1-5",
torch_dtype=torch.float16
).to("cuda")
pipe.unet.load_attn_procs("your_username_or_org/lora-hyundai-sd15/checkpoint-XXX")
prompt = "Front of SUV"
image = pipe(prompt).images[0]
image.save("output.png")
- Downloads last month
- 2
Inference Providers
NEW
This model isn't deployed by any Inference Provider.
π
Ask for provider support
Model tree for mingyu-oo/stable-diffusion-v1-5-lora-hyundai-car-design
Base model
stable-diffusion-v1-5/stable-diffusion-v1-5