app.py

import torch

from datasets import load_dataset
import transformers
from diffusers import StableDiffusionPipeline

import gradio as gr

from random import randrange
import os

MY_SECRET_TOKEN = os.environ.get('stable-diffusion')

data = load_dataset("mfumanelli/movies-small")
data = data['train'].to_pandas()

model_id = 'CompVis/stable-diffusion-v1-4'
device = torch.device('cpu' if not torch.cuda.is_available() else 'cuda')

pipe = StableDiffusionPipeline.from_pretrained(model_id, use_auth_token=MY_SECRET_TOKEN, revision='fp16')
pipe = pipe.to(device)


def infer(prompt, samples, steps, scale):
    generator = torch.Generator(device=device)

    if device.type == 'cuda':
        with torch.autocast(device.type):
            images_list = pipe(
                [prompt] * samples,
                num_inference_steps=steps,
                guidance_scale=scale,
                generator=generator,
            )
    else:
        images_list = pipe(
            [prompt] * samples,
            num_inference_steps=steps,
            guidance_scale=scale,
            generator=generator,
        )

    return images_list


def generate_movie():
    seed = randrange(data.shape[0])
    plot = data.iloc[seed]["plot_synopsis_sum"]
    image = infer(plot, 1, 50, 7.5)

    return image["sample"][0], seed


def movie_title(seed):
    return data.iloc[int(seed)]["title"]


css = """
        .gradio-container {
            font-family: 'IBM Plex Sans', sans-serif;
        }
        .gr-button {
            color: white;
            border-color: black;
            background: black;
        }
        input[type='range'] {
            accent-color: black;
        }
        .dark input[type='range'] {
            accent-color: #dfdfdf;
        }
        .container {
            max-width: 730px;
            margin: auto;
            padding-top: 1.5rem;
        }
        #iamge {
            min-height: 22rem;
            margin-bottom: 15px;
            margin-left: auto;
            margin-right: auto;
            border-bottom-right-radius: .5rem !important;
            border-bottom-left-radius: .5rem !important;
        }
        #iamge>div>.h-full {
            min-height: 20rem;
        }
        .details:hover {
            text-decoration: underline;
        }
        .gr-button {
            white-space: nowrap;
        }
        .gr-button:focus {
            border-color: rgb(147 197 253 / var(--tw-border-opacity));
            outline: none;
            box-shadow: var(--tw-ring-offset-shadow), var(--tw-ring-shadow), var(--tw-shadow, 0 0 #0000);
            --tw-border-opacity: 1;
            --tw-ring-offset-shadow: var(--tw-ring-inset) 0 0 0 var(--tw-ring-offset-width) var(--tw-ring-offset-color);
            --tw-ring-shadow: var(--tw-ring-inset) 0 0 0 calc(3px var(--tw-ring-offset-width)) var(--tw-ring-color);
            --tw-ring-color: rgb(191 219 254 / var(--tw-ring-opacity));
            --tw-ring-opacity: .5;
        }
        .footer {
            margin-bottom: 45px;
            margin-top: 35px;
            text-align: center;
            border-bottom: 1px solid #e5e5e5;
        }
        .footer>p {
            font-size: .8rem;
            display: inline-block;
            padding: 0 10px;
            transform: translateY(10px);
            background: white;
        }
        .dark .footer {
            border-color: #303030;
        }
        .dark .footer>p {
            background: #0b0f19;
        }
        .acknowledgments h4{
            margin: 1.25em 0 .25em 0;
            font-weight: bold;
            font-size: 115%;
        }
"""

with gr.Blocks(css=css) as demo:
    gr.HTML(
        """
            <div style="text-align: center; max-width: 650px; margin: 0 auto;">
              <div
                style="
                  display: inline-flex;
                  align-items: center;
                  gap: 0.8rem;
                  font-size: 1.75rem;
                "
              >

                <svg style="color: red" xmlns="http://www.w3.org/2000/svg" viewBox="0 0 256 256", height="0.85em" width="0.85em">
                <rect width="18em" height="18em" fill="none"></rect>
                <path d="M128,216S28,160,28,92A52,52,0,0,1,128,72h0A52,52,0,0,1,228,92C228,160,128,216,128,216Z" fill="#d63e25" stroke="#d63e25" stroke-linecap="round"
                stroke-linejoin="round" stroke-width="12"></path></svg>
                <h1 style="font-weight: 900; margin-bottom: 7px;">
                  Stable Diffusion Loves Cinema
                </h1>
              </div>
              <p style="margin-bottom: 20px; font-size: 94%">
                Stable Diffusion is a state-of-the-art text-to-image model that generates images from text, 
                in this demo it is used to generate movie scenes from their storyline. <br></p>
                <hr style="height:2px;border-width:0;color:gray;background-color:gray">
                <br>
              <p align="left" style="margin-bottom: 10px; font-size: 94%">
                <b>Instructions</b>: press the "Generate a movie scene!" button to generate an image and try to see if you can guess the movie.
                You can see if you guessed right by pressing the "Tell me the title" button.

              </p>
            </div>
        """
    )
    with gr.Group():
        with gr.Box():
            with gr.Row().style(mobile_collapse=False, equal_height=True):
                b1 = gr.Button("Generate a movie scene!").style(
                    margin=False,
                    rounded=(False, True, True, False),
                )
                b2 = gr.Button("Tell me the title").style(
                    margin=False,
                    rounded=(False, True, True, False),
                )
            text = gr.Textbox(label="Title:")
            image = gr.Image(
                label="Generated images", show_label=False, elem_id="image"
            ).style(height="auto")

            seed = gr.Number(visible=False)

            b1.click(generate_movie, inputs=None, outputs=[image, seed])
            b2.click(movie_title, inputs=seed, outputs=text)

demo.launch()