File size: 2,248 Bytes
7a5cf75 2d709ef 7a5cf75 2d709ef |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 |
import gradio as gr
import torch
from diffusers import DiffusionPipeline
# Carregar o modelo base
base_model = "stabilityai/stable-diffusion-2-1"
pipe = DiffusionPipeline.from_pretrained(base_model, torch_dtype=torch.bfloat16)
# Carregar o LoRA a partir do repositório fornecido
lora_repo = "Shakker-Labs/FLUX.1-dev-LoRA-blended-realistic-illustration"
pipe.load_lora_weights(lora_repo)
pipe.to("cuda")
MAX_SEED = 2**32 - 1
def generate_image(prompt, cfg_scale, steps, randomize_seed, seed, width, height, lora_scale):
if randomize_seed:
seed = random.randint(0, MAX_SEED)
generator = torch.Generator(device="cuda").manual_seed(seed)
image = pipe(
prompt=prompt,
num_inference_steps=steps,
guidance_scale=cfg_scale,
width=width,
height=height,
generator=generator,
joint_attention_kwargs={"scale": lora_scale},
).images[0]
return image, seed
with gr.Blocks() as app:
gr.Markdown("# Flux RealismLora Image Generator")
with gr.Row():
with gr.Column(scale=3):
prompt = gr.TextArea(label="Prompt", placeholder="Digite o prompt", lines=5)
cfg_scale = gr.Slider(label="CFG Scale", minimum=1, máximo=20, passo=0.5, valor=7.5)
steps = gr.Slider(label="Steps", mínimo=1, máximo=100, passo=1, valor=50)
width = gr.Slider(label="Width", mínimo=256, máximo=1536, passo=64, valor=768)
height = gr.Slider(label="Height", mínimo=256, máximo=1536, passo=64, valor=768)
randomize_seed = gr.Checkbox(False, label="Randomize seed")
seed = gr.Slider(label="Seed", mínimo=0, máximo=MAX_SEED, passo=1, valor=42)
lora_scale = gr.Slider(label="LoRA Scale", mínimo=0, máximo=1, passo=0.01, valor=0.85)
generate_button = gr.Button("Generate")
with gr.Column(scale=1):
result = gr.Image(label="Generated Image")
gr.Markdown("Gere imagens usando RealismLora com um prompt de texto.")
generate_button.click(
generate_image,
inputs=[prompt, cfg_scale, steps, randomize_seed, seed, width, height, lora_scale],
outputs=[result, seed]
)
app.queue()
app.launch() |