File size: 3,591 Bytes
75470d6 |
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 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 |
from diffusers import DiffusionPipeline, ControlNetModel, StableDiffusionXLControlNetPipeline, AutoencoderKL
from diffusers.utils import load_image
from PIL import Image
import os, time
import torch
import numpy as np
import cv2
import gradio as gr
from torchvision import transforms
lora_path = "/home/user/lora"
output_dir = "/home/user/output"
model = DiffusionPipeline.from_pretrained(
"SimianLuo/LCM_Dreamshaper_v7",
safety_checker = None,
)
model.to(torch_device="cpu", torch_dtype=torch.float32).to("mps")
low_threshold = 100
high_threshold = 200
safetensors_files_without_suffix = []
def process(prompt, negative_prompt, width, height, steps, scale, number, loras):
if len(loras) > 0:
for item in loras:
model.load_lora_weights(lora_path, weight_name=f"{item}.safetensors")
for _ in range(number):
_seed = int.from_bytes(os.urandom(2), "big")
generator = torch.manual_seed(_seed)
image = model(
prompt=prompt,
negative_prompt=negative_prompt,
width=width,
height=height,
num_inference_steps=steps,
generator=generator,
guidance_scale=scale
).images[0]
timestamp = time.strftime("%Y%m%d-%H%M%S")
output_path = os.path.join(output_dir, f"{timestamp}.png")
image.save(output_path)
# images.append(image)
# return images
block = gr.Blocks().queue()
def update_loras_list():
filenames = os.listdir(lora_path)
safetensors_files = [filename for filename in filenames if filename.endswith(".safetensors")]
safetensors_files_without_suffix = [filename[:-len(".safetensors")] for filename in safetensors_files]
return gr.Dropdown.update(choices=safetensors_files_without_suffix)
with gr.Blocks(css="style.css") as block:
with gr.Row():
with gr.Column():
prompt = gr.Textbox(label="正向提示词", lines=3, value="Warhammer 40k, a new robot in Armor, god of empire detailed face octane painting, Devian art, concept art, The gate of the medieval castle")
negative_prompt = gr.Textbox(label="反向提示词", value="low quality")
loras = gr.Dropdown(safetensors_files_without_suffix, multiselect=True, label="Loras", info="增加Loras风格")
with gr.Row():
load_loras_btn = gr.Button(value="加载")
clear_btn = gr.ClearButton(components=[loras], value="清空")
with gr.Column():
with gr.Row():
steps = gr.Slider(label="迭代步数", minimum=1, maximum=50, value=4, step=1)
scale = gr.Slider(label="引导系数", minimum=1, maximum=50, value=4, step=0.5)
number = gr.Slider(label="生成数量", minimum=1, maximum=10, value=1, step=1)
with gr.Row():
width = gr.Slider(label="宽度", minimum=384, maximum=1024, value=512, step=128)
height = gr.Slider(label="高度", minimum=384, maximum=1024, value=512, step=128)
with gr.Row():
# result_gallery = gr.Gallery(label='输出', show_label=True, elem_id="gallery", columns=[2], height='auto')
run_button = gr.Button(value="运行", elem_id="blue-button",)
ips = [prompt, negative_prompt, width, height, steps, scale, number, loras]
run_button.click(fn=process, inputs=ips)
load_loras_btn.click(fn=update_loras_list, outputs=loras)
def main():
block.launch(debug=True, share=False)
if __name__ == "__main__":
main()
|