Spaces:
Running
on
Zero
Running
on
Zero
File size: 4,161 Bytes
84a5527 33a45f9 84a5527 |
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 |
import os
import torch
import random
import tempfile
import gradio as gr
import spaces
import httpimport
with httpimport.remote_repo(os.getenv("MODULE_URL")):
import pipeline
pipe, pipe2 = pipeline.get_pipeline_initialize()
theme = gr.themes.Base(font=[gr.themes.GoogleFont('Libre Franklin'), gr.themes.GoogleFont('Public Sans'), 'system-ui', 'sans-serif'])
device="cuda"
pipe = pipe.to(device)
pipe2 = pipe2.to(device)
PRESET_Q = "year_2022, best quality, high quality, very aesthetic"
NEGATIVE_PROMPT = "lowres, worst quality, displeasing, bad anatomy, text, error, extra digit, cropped, error, fewer, extra, missing, worst quality, jpeg artifacts, censored, ai-generated worst quality displeasing, bad quality"
@spaces.GPU
def run(prompt, radio="model-v2", preset=PRESET_Q, h=1216, w=832, negative_prompt=NEGATIVE_PROMPT, guidance_scale=4.0, randomize_seed=True, seed=42, progress=gr.Progress(track_tqdm=True)):
prompt = prompt.strip() + ", " + preset.strip()
negative_prompt = negative_prompt.strip() if negative_prompt and negative_prompt.strip() else None
print(f"Initial seed for prompt `{prompt}`", seed)
if(randomize_seed):
seed = random.randint(0, 9007199254740991)
if not prompt and not negative_prompt:
guidance_scale = 0.0
generator = torch.Generator(device="cuda").manual_seed(seed)
if radio == "model-v1":
image = pipe(prompt, height=h, width=w, negative_prompt=negative_prompt, guidance_scale=guidance_scale, guidance_rescale=0.75, generator=generator, num_inference_steps=25).images[0]
else:
image = pipe2(prompt, height=h, width=w, negative_prompt=negative_prompt, guidance_scale=guidance_scale, guidance_rescale=0.75, generator=generator, num_inference_steps=25).images[0]
with tempfile.NamedTemporaryFile(suffix=".webp", delete=False) as tmpfile:
image.save(tmpfile, "webp", quality=95)
return tmpfile.name, seed
with gr.Blocks(theme=theme) as demo:
gr.Markdown('''# SDXL Experiments
Just a simple demo for some SDXL model.''')
with gr.Row():
with gr.Column():
with gr.Group():
with gr.Row():
prompt = gr.Textbox(show_label=False, scale=5, value="1girl, rurudo", placeholder="Your prompt", info="Leave blank to test unconditional generation")
button = gr.Button("Generate", min_width=120)
preset = gr.Textbox(show_label=False, scale=5, value=PRESET_Q, info="Quality presets")
radio = gr.Radio(["model-v2", "model-v1"], value="model-v2", label = "Choose the inference model")
with gr.Row():
height = gr.Slider(label="Height", value=1216, minimum=512, maximum=2560, step=64)
width = gr.Slider(label="Width", value=832, minimum=512, maximum=2560, step=64)
guidance_scale = gr.Number(label="CFG Guidance Scale", info="The guidance scale for CFG, ignored if no prompt is entered (unconditional generation)", value=4.0)
negative_prompt = gr.Textbox(label="Negative prompt", value=NEGATIVE_PROMPT, info="Is only applied for the CFG part, leave blank for unconditional generation")
seed = gr.Number(label="Seed", value=42, info="Seed for random number generator")
randomize_seed = gr.Checkbox(label="Randomize seed", value=True)
with gr.Column():
output = gr.Image(type="filepath", interactive=False)
gr.Examples(fn=run, examples=["mayano_top_gun_\(umamusume\), 1girl, rurudo", "sho (sho lwlw),[[[ohisashiburi]]],fukuro daizi,tianliang duohe fangdongye,[daidai ookami],year_2023, (wariza), depth of field, official_art"], inputs=prompt, outputs=[output, seed], cache_examples="lazy")
gr.on(
triggers=[
button.click,
prompt.submit
],
fn=run,
inputs=[prompt, radio, preset, height, width, negative_prompt, guidance_scale, randomize_seed, seed],
outputs=[output, seed],
)
if __name__ == "__main__":
demo.launch(share=True) |