GlyphControl / app.py
yyk19's picture
transfer the ema checkpoint to the clean version.
bc1f1f4
raw
history blame
9.55 kB
from cldm.ddim_hacked import DDIMSampler
import math
from omegaconf import OmegaConf
from scripts.rendertext_tool import Render_Text, load_model_from_config
import gradio as gr
def process_multi_wrapper(rendered_txt_0, rendered_txt_1, rendered_txt_2, rendered_txt_3,
shared_prompt,
width_0, width_1, width_2, width_3,
ratio_0, ratio_1, ratio_2, ratio_3,
top_left_x_0, top_left_x_1, top_left_x_2, top_left_x_3,
top_left_y_0, top_left_y_1, top_left_y_2, top_left_y_3,
yaw_0, yaw_1, yaw_2, yaw_3,
num_rows_0, num_rows_1, num_rows_2, num_rows_3,
shared_num_samples, shared_image_resolution,
shared_ddim_steps, shared_guess_mode,
shared_strength, shared_scale, shared_seed,
shared_eta, shared_a_prompt, shared_n_prompt):
rendered_txt_values = [rendered_txt_0, rendered_txt_1, rendered_txt_2, rendered_txt_3]
width_values = [width_0, width_1, width_2, width_3]
ratio_values = [ratio_0, ratio_1, ratio_2, ratio_3]
top_left_x_values = [top_left_x_0, top_left_x_1, top_left_x_2, top_left_x_3]
top_left_y_values = [top_left_y_0, top_left_y_1, top_left_y_2, top_left_y_3]
yaw_values = [yaw_0, yaw_1, yaw_2, yaw_3]
num_rows_values = [num_rows_0, num_rows_1, num_rows_2, num_rows_3]
return render_tool.process_multi(rendered_txt_values, shared_prompt,
width_values, ratio_values,
top_left_x_values, top_left_y_values,
yaw_values, num_rows_values,
shared_num_samples, shared_image_resolution,
shared_ddim_steps, shared_guess_mode,
shared_strength, shared_scale, shared_seed,
shared_eta, shared_a_prompt, shared_n_prompt
)
def process_multi_wrapper_only_show_rendered(rendered_txt_0, rendered_txt_1, rendered_txt_2, rendered_txt_3,
shared_prompt,
width_0, width_1, width_2, width_3,
ratio_0, ratio_1, ratio_2, ratio_3,
top_left_x_0, top_left_x_1, top_left_x_2, top_left_x_3,
top_left_y_0, top_left_y_1, top_left_y_2, top_left_y_3,
yaw_0, yaw_1, yaw_2, yaw_3,
num_rows_0, num_rows_1, num_rows_2, num_rows_3,
shared_num_samples, shared_image_resolution,
shared_ddim_steps, shared_guess_mode,
shared_strength, shared_scale, shared_seed,
shared_eta, shared_a_prompt, shared_n_prompt):
rendered_txt_values = [rendered_txt_0, rendered_txt_1, rendered_txt_2, rendered_txt_3]
width_values = [width_0, width_1, width_2, width_3]
ratio_values = [ratio_0, ratio_1, ratio_2, ratio_3]
top_left_x_values = [top_left_x_0, top_left_x_1, top_left_x_2, top_left_x_3]
top_left_y_values = [top_left_y_0, top_left_y_1, top_left_y_2, top_left_y_3]
yaw_values = [yaw_0, yaw_1, yaw_2, yaw_3]
num_rows_values = [num_rows_0, num_rows_1, num_rows_2, num_rows_3]
return render_tool.process_multi(rendered_txt_values, shared_prompt,
width_values, ratio_values,
top_left_x_values, top_left_y_values,
yaw_values, num_rows_values,
shared_num_samples, shared_image_resolution,
shared_ddim_steps, shared_guess_mode,
shared_strength, shared_scale, shared_seed,
shared_eta, shared_a_prompt, shared_n_prompt,
only_show_rendered_image=True)
# cfg = OmegaConf.load("config.yaml")
# model = load_model_from_config(cfg, "model_states.pt", verbose=True)
cfg = OmegaConf.load("config.yaml")
model = load_model_from_config(cfg, "model.ckpt", verbose=True)
ddim_sampler = DDIMSampler(model)
render_tool = Render_Text(model)
block = gr.Blocks().queue()
with block:
with gr.Row():
gr.Markdown("## Control Stable Diffusion with Glyph Images")
only_show_rendered_image = gr.Number(value=1, visible=False)
with gr.Column():
with gr.Row():
for i in range(4):
with gr.Column():
exec(f"""rendered_txt_{i} = gr.Textbox(label=f"rendered_txt {i+1}")""")
with gr.Accordion(f"Advanced options {i+1}", open=False):
exec(f"""width_{i} = gr.Slider(label="bbox_width", minimum=0., maximum=1, value=0.3, step=0.01) """)
exec(f"""ratio_{i} = gr.Slider(label="bbox_width_height_ratio", minimum=0., maximum=5, value=0., step=0.02) """)
exec(f"""top_left_x_{i} = gr.Slider(label="bbox_top_left_x", minimum=0., maximum=1, value={0.35 - 0.25 * math.cos(math.pi * i)}, step=0.01) """)
exec(f"""top_left_y_{i} = gr.Slider(label="bbox_top_left_y", minimum=0., maximum=1, value={0.1 if i < 2 else 0.6}, step=0.01) """)
exec(f"""yaw_{i} = gr.Slider(label="bbox_yaw", minimum=-180, maximum=180, value=0, step=5) """)
exec(f"""num_rows_{i} = gr.Slider(label="num_rows", minimum=1, maximum=4, value=1, step=1) """)
with gr.Row():
with gr.Column():
shared_prompt = gr.Textbox(label="Shared Prompt")
with gr.Row():
run_button = gr.Button(value="Run")
show_render_button = gr.Button(value="Only Rendered")
with gr.Accordion("Shared Advanced options", open=False):
shared_num_samples = gr.Slider(label="Images", minimum=1, maximum=12, value=1, step=1)
shared_image_resolution = gr.Slider(label="Image Resolution", minimum=256, maximum=768, value=512, step=64)
shared_strength = gr.Slider(label="Control Strength", minimum=0.0, maximum=2.0, value=1.0, step=0.01)
shared_guess_mode = gr.Checkbox(label='Guess Mode', value=False)
shared_scale = gr.Slider(label="Guidance Scale", minimum=0.1, maximum=30.0, value=9.0, step=0.1)
shared_ddim_steps = gr.Slider(label="Steps", minimum=1, maximum=100, value=20, step=1)
shared_seed = gr.Slider(label="Seed", minimum=-1, maximum=2147483647, step=1, randomize=True)
shared_eta = gr.Number(label="eta (DDIM)", value=0.0)
shared_a_prompt = gr.Textbox(label="Added Prompt", value='best quality, extremely detailed')
shared_n_prompt = gr.Textbox(label="Negative Prompt",
value='longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality')
with gr.Row():
result_gallery = gr.Gallery(label='Output', show_label=False, elem_id="gallery").style(grid=2, height='auto')
run_button.click(fn=process_multi_wrapper,
inputs=[rendered_txt_0, rendered_txt_1, rendered_txt_2, rendered_txt_3,
shared_prompt,
width_0, width_1, width_2, width_3,
ratio_0, ratio_1, ratio_2, ratio_3,
top_left_x_0, top_left_x_1, top_left_x_2, top_left_x_3,
top_left_y_0, top_left_y_1, top_left_y_2, top_left_y_3,
yaw_0, yaw_1, yaw_2, yaw_3,
num_rows_0, num_rows_1, num_rows_2, num_rows_3,
shared_num_samples, shared_image_resolution,
shared_ddim_steps, shared_guess_mode,
shared_strength, shared_scale, shared_seed,
shared_eta, shared_a_prompt, shared_n_prompt],
outputs=[result_gallery])
show_render_button.click(fn=process_multi_wrapper_only_show_rendered,
inputs=[rendered_txt_0, rendered_txt_1, rendered_txt_2, rendered_txt_3,
shared_prompt,
width_0, width_1, width_2, width_3,
ratio_0, ratio_1, ratio_2, ratio_3,
top_left_x_0, top_left_x_1, top_left_x_2, top_left_x_3,
top_left_y_0, top_left_y_1, top_left_y_2, top_left_y_3,
yaw_0, yaw_1, yaw_2, yaw_3,
num_rows_0, num_rows_1, num_rows_2, num_rows_3,
shared_num_samples, shared_image_resolution,
shared_ddim_steps, shared_guess_mode,
shared_strength, shared_scale, shared_seed,
shared_eta, shared_a_prompt, shared_n_prompt],
outputs=[result_gallery])
block.launch()