Spaces:
Running
on
Zero
Running
on
Zero
import gradio as gr | |
import numpy as np | |
import random | |
import spaces | |
import torch | |
from diffusers import DiffusionPipeline | |
from transformers import pipeline | |
# λ²μ νμ΄νλΌμΈ λ° νλμ¨μ΄ μ€μ | |
device = "cuda" if torch.cuda.is_available() else "cpu" | |
translator = pipeline("translation", model="Helsinki-NLP/opus-mt-ko-en", device=device) | |
dtype = torch.bfloat16 | |
pipe = DiffusionPipeline.from_pretrained("black-forest-labs/FLUX.1-schnell", torch_dtype=dtype).to(device) | |
MAX_SEED = np.iinfo(np.int32).max | |
MAX_IMAGE_SIZE = 2048 | |
def infer(prompt, seed=42, randomize_seed=False, width=1024, height=1024, num_inference_steps=4, progress=gr.Progress(track_tqdm=True)): | |
if randomize_seed: | |
seed = random.randint(0, MAX_SEED) | |
generator = torch.Generator().manual_seed(seed) | |
# νκΈ μ λ ₯ κ°μ§ λ° λ²μ | |
if any('\uAC00' <= char <= '\uD7A3' for char in prompt): | |
print("νκ΅μ΄ ν둬ννΈ λ²μ μ€...") | |
translated_prompt = translator(prompt, max_length=512)[0]['translation_text'] | |
print("λ²μλ ν둬ννΈ:", translated_prompt) | |
prompt = translated_prompt | |
image = pipe( | |
prompt = prompt, | |
width = width, | |
height = height, | |
num_inference_steps = num_inference_steps, | |
generator = generator, | |
guidance_scale=0.0 | |
).images[0] | |
return image, seed | |
examples = [ | |
["[μμ: νλμ] [λμμΈ μ»¨μ : λ‘μΌ] [ν μ€νΈ: 'μΈκ³'] [λ°°κ²½: νλμ]μ μλ‘μ΄ λ‘κ³ λ§λ€κΈ°"], | |
["[μμ: νλμ] [λμμΈ μ»¨μ : μ°μ£Ό] [ν μ€νΈ: 'μ½μΉ΄μ½λΌ'] [λ°°κ²½: λ€μ±λ‘μ΄ μμ]μ μλ‘μ΄ λ‘κ³ λ§λ€κΈ°"], | |
["λ°©ν¨ μμ μλ κ°λ¨ν λ―Έλμ μΈ μΉ΄λ―ΈμΉ΄μ λλ‘ λ‘κ³ , λ―Έλλ©λ¦¬μ€ν±, 벑ν°, 2D, λ¨μν μ , ν°μ λ°°κ²½ --v 4"], | |
["[μμ: νλμ] [λμμΈ μ»¨μ : μ°] [ν μ€νΈ: 'abc@gmail.com'] [λ°°κ²½: λΉ¨κ°μ]μ μλ‘μ΄ λ‘κ³ λ§λ€κΈ°"], | |
["[μμ: νλμ] [λμμΈ μ»¨μ : μ¬λ] [ν μ€νΈ: 'ABC.COM'] [λ°°κ²½: λ Έλμ]μ μλ‘μ΄ λ‘κ³ λ§λ€κΈ°"], | |
["[μμ: νλμ] [λμμΈ μ»¨μ : μ§] [ν μ€νΈ: 'T.010-1234-1234'] [λ°°κ²½: λ€μ±λ‘μ΄ μμ]μ μλ‘μ΄ λ‘κ³ λ§λ€κΈ°"], | |
["[μμ: νλμ] [λμμΈ μ»¨μ : μ¬μ] [ν μ€νΈ: 'μΆκ΅¬ ν΄λ½'] [λ°°κ²½: μ΄λ‘μ]μ μλ‘μ΄ λ‘κ³ λ§λ€κΈ°"] | |
] | |
css = """ | |
footer { | |
visibility: hidden; | |
} | |
""" | |
with gr.Blocks(theme="Nymbo/Nymbo_Theme", css=css) as demo: | |
with gr.Column(elem_id="col-container"): | |
with gr.Row(): | |
prompt = gr.Text( | |
label="ν둬ννΈ", | |
show_label=False, | |
max_lines=1, | |
placeholder="ν둬ννΈλ₯Ό μ λ ₯νμΈμ", | |
container=False, | |
elem_id="prompt" | |
) | |
run_button = gr.Button("μ€ν", scale=0) | |
result = gr.Image(label="κ²°κ³Ό", show_label=False, elem_id="result") | |
with gr.Accordion("κ³ κΈ μ€μ ", open=False, elem_id="advanced-settings"): | |
seed = gr.Slider( | |
label="μλ", | |
minimum=0, | |
maximum=MAX_SEED, | |
step=1, | |
value=0, | |
) | |
randomize_seed = gr.Checkbox(label="μλ 무μμν", value=True) | |
with gr.Row(): | |
width = gr.Slider( | |
label="λλΉ", | |
minimum=256, | |
maximum=MAX_IMAGE_SIZE, | |
step=32, | |
value=512, | |
) | |
height = gr.Slider( | |
label="λμ΄", | |
minimum=256, | |
maximum=MAX_IMAGE_SIZE, | |
step=32, | |
value=512, | |
) | |
with gr.Row(): | |
num_inference_steps = gr.Slider( | |
label="μΆλ‘ λ¨κ³ μ", | |
minimum=1, | |
maximum=50, | |
step=1, | |
value=4, | |
) | |
gr.Examples( | |
examples=examples, | |
fn=infer, | |
inputs=[prompt], | |
outputs=[result, seed], | |
cache_examples="lazy" | |
) | |
gr.on( | |
triggers=[run_button.click, prompt.submit], | |
fn=infer, | |
inputs=[prompt, seed, randomize_seed, width, height, num_inference_steps], | |
outputs=[result, seed] | |
) | |
demo.launch() |