app.py

#!/usr/bin/env python

from __future__ import annotations

import argparse

import gradio as gr

from model import Model

TITLE = '# Anime Face Generation with [Diffusers](https://github.com/huggingface/diffusers)'
DESCRIPTION = 'Expected execution time on Hugging Face Spaces: 5s (DDIM, 20 steps), 6s (PNDM, 20 steps), 247s (DDPM, 1000 steps)'
FOOTER = '<img id="visitor-badge" src="https://visitor-badge.glitch.me/badge?page_id=hysts.diffusers-anime-faces" alt="visitor badge" />'


def create_simple_demo(model: Model) -> gr.Blocks:
    with gr.Blocks() as demo:
        run_button = gr.Button('Generate')
        with gr.Tabs():
            with gr.TabItem('Result (Superresolved)'):
                result = gr.Image(show_label=False, elem_id='result-grid')
            with gr.TabItem('Result (Raw)'):
                result_raw = gr.Image(show_label=False,
                                      elem_id='result-grid-raw')
        run_button.click(fn=model.run_simple,
                         inputs=None,
                         outputs=[result, result_raw])
    return demo


def create_advanced_demo(model: Model) -> gr.Blocks:
    def update_num_steps(name: str) -> dict:
        visible = name != 'DDPM'
        if name == 'PNDM':
            minimum = 4
            maximum = 100
        else:
            minimum = 1
            maximum = 200
        return gr.Slider.update(visible=visible,
                                minimum=minimum,
                                maximum=maximum,
                                value=20)

    with gr.Blocks() as demo:
        gr.Markdown(DESCRIPTION)

        with gr.Row():
            with gr.Column():
                with gr.Group():
                    model_name = gr.Dropdown(model.MODEL_NAMES,
                                             value=model.MODEL_NAMES[0],
                                             label='Model',
                                             interactive=False)
                    scheduler_type = gr.Radio(choices=['DDPM', 'DDIM', 'PNDM'],
                                              value='DDIM',
                                              label='Scheduler')
                    num_steps = gr.Slider(1,
                                          200,
                                          step=1,
                                          value=20,
                                          label='Number of Steps')
                    randomize_seed = gr.Checkbox(value=False,
                                                 label='Randomize Seed')
                    seed = gr.Slider(0,
                                     100000,
                                     step=1,
                                     value=1234,
                                     label='Seed')
                    run_button = gr.Button('Run')
            with gr.Column():
                with gr.Tabs():
                    with gr.TabItem('Result (Superresolved)'):
                        result = gr.Image(show_label=False, elem_id='result')
                    with gr.TabItem('Result (Raw)'):
                        result_raw = gr.Image(show_label=False,
                                              elem_id='result-raw')
                    with gr.TabItem('Denoising Process'):
                        result_video = gr.Video(show_label=False,
                                                elem_id='result-video')

        scheduler_type.change(fn=update_num_steps,
                              inputs=scheduler_type,
                              outputs=num_steps,
                              queue=False)
        run_button.click(fn=model.run,
                         inputs=[
                             model_name,
                             scheduler_type,
                             num_steps,
                             randomize_seed,
                             seed,
                         ],
                         outputs=[
                             result,
                             result_raw,
                             seed,
                             result_video,
                         ])
    return demo


def create_sample_image_view_demo() -> gr.Blocks:
    def get_sample_image_url(file_name: str) -> str:
        sample_image_dir = 'https://huggingface.co/spaces/hysts/diffusers-anime-faces/resolve/main/samples'
        return f'{sample_image_dir}/{file_name}'

    def get_sample_image_markdown(name: str) -> str:
        model_name = name.split()[0]
        if name == 'ddpm-128-exp000 (DDPM)':
            scheduler = 'DDPM'
            steps = 1000
            file_name = f'{model_name}.png'
        elif name == 'ddpm-128-exp000 (DDIM, 20 steps)':
            scheduler = 'DDIM'
            steps = 20
            file_name = f'{model_name}-ddim-20steps.png'
        else:
            raise ValueError
        url = get_sample_image_url(file_name)
        text = f'''
                - size: 128x128
                - seed: 0-99
                - scheduler: {scheduler}
                - steps: {steps}

                ![sample images]({url})'''
        return text

    with gr.Blocks() as demo:
        with gr.Row():
            model_name = gr.Dropdown([
                'ddpm-128-exp000 (DDPM)',
                'ddpm-128-exp000 (DDIM, 20 steps)',
            ],
                                     value='ddpm-128-exp000 (DDPM)',
                                     label='Model')
        with gr.Row():
            text = get_sample_image_markdown(model_name.value)
            sample_images = gr.Markdown(text)

        model_name.change(fn=get_sample_image_markdown,
                          inputs=model_name,
                          outputs=sample_images)
    return demo


def main():
    parser = argparse.ArgumentParser()
    parser.add_argument('--device', type=str, default='cpu')
    args = parser.parse_args()
    model = Model(args.device)

    with gr.Blocks(css='style.css') as demo:
        gr.Markdown(TITLE)
        with gr.Tabs():
            with gr.TabItem('Simple Mode'):
                create_simple_demo(model)
            with gr.TabItem('Advanced Mode'):
                create_advanced_demo(model)
            with gr.TabItem('Sample Images'):
                create_sample_image_view_demo()
        gr.Markdown(FOOTER)
    demo.launch(enable_queue=True, share=False)


if __name__ == '__main__':
    main()