app.py

from diffusers import AutoencoderKL, UNet2DConditionModel, StableDiffusionPipeline, StableDiffusionImg2ImgPipeline, DPMSolverMultistepScheduler
import gradio as gr
import torch
from PIL import Image
import utils

is_colab = utils.is_google_colab()


class Model:
    def __init__(self, name, path, prefix):
        self.name = name
        self.path = path
        self.prefix = prefix
        self.pipe_t2i = None
        self.pipe_i2i = None


models = [
    Model("Beeple", "riccardogiorato/beeple-diffusion", "beeple style "),
    Model("Avatar", "riccardogiorato/avatar-diffusion", "avatartwow style "),
    Model("Beksinski", "s3nh/beksinski-style-stable-diffusion", "beksinski style "),
    Model("Poolsuite", "prompthero/poolsuite", "poolsuite style "),
    Model("Robo Diffusion", "nousr/robo-diffusion", ""),
    Model("Guohua", "Langboat/Guohua-Diffusion", "guohua style "),
    Model("JWST", "dallinmackay/JWST-Deep-Space-diffusion", "JWST ")
]

scheduler = DPMSolverMultistepScheduler(
    beta_start=0.00085,
    beta_end=0.012,
    beta_schedule="scaled_linear",
    num_train_timesteps=1000,
    trained_betas=None,
    predict_epsilon=True,
    thresholding=False,
    algorithm_type="dpmsolver++",
    solver_type="midpoint",
    lower_order_final=True,
)

last_mode = "txt2img"
current_model = models[0]
current_model_path = current_model.path

if is_colab:
    pipe = StableDiffusionPipeline.from_pretrained(
        current_model.path, torch_dtype=torch.float16, scheduler=scheduler)

else:  # download all models
    vae = AutoencoderKL.from_pretrained(
        current_model.path, subfolder="vae", torch_dtype=torch.float16)
    for model in models:
        try:
            unet = UNet2DConditionModel.from_pretrained(
                model.path, subfolder="unet", torch_dtype=torch.float16)
            model.pipe_t2i = StableDiffusionPipeline.from_pretrained(
                model.path, unet=unet, vae=vae, torch_dtype=torch.float16, scheduler=scheduler)
            model.pipe_i2i = StableDiffusionImg2ImgPipeline.from_pretrained(
                model.path, unet=unet, vae=vae, torch_dtype=torch.float16, scheduler=scheduler)
        except:
            models.remove(model)
    pipe = models[0].pipe_t2i

if torch.cuda.is_available():
    pipe = pipe.to("cuda")

device = "GPU 🔥" if torch.cuda.is_available() else "CPU 🥶"


def inference(model_name, prompt, guidance, steps, width=512, height=512, seed=0, img=None, strength=0.5, neg_prompt=""):

    global current_model
    for model in models:
        if model.name == model_name:
            current_model = model
            model_path = current_model.path

    generator = torch.Generator('cuda').manual_seed(
        seed) if seed != 0 else None

    if img is not None:
        return img_to_img(model_path, prompt, neg_prompt, img, strength, guidance, steps, width, height, generator)
    else:
        return txt_to_img(model_path, prompt, neg_prompt, guidance, steps, width, height, generator)


def txt_to_img(model_path, prompt, neg_prompt, guidance, steps, width, height, generator=None):

    global last_mode
    global pipe
    global current_model_path
    if model_path != current_model_path or last_mode != "txt2img":
        current_model_path = model_path

        if is_colab:
            pipe = StableDiffusionPipeline.from_pretrained(
                current_model_path, torch_dtype=torch.float16, scheduler=scheduler)
        else:
            pipe.to("cpu")
            pipe = current_model.pipe_t2i

        if torch.cuda.is_available():
            pipe = pipe.to("cuda")
        last_mode = "txt2img"

    prompt = current_model.prefix + prompt
    result = pipe(
        prompt,
        negative_prompt=neg_prompt,
        # num_images_per_prompt=n_images,
        num_inference_steps=int(steps),
        guidance_scale=guidance,
        width=width,
        height=height,
        generator=generator)

    return replace_nsfw_images(result)


def img_to_img(model_path, prompt, neg_prompt, img, strength, guidance, steps, width, height, generator=None):

    global last_mode
    global pipe
    global current_model_path
    if model_path != current_model_path or last_mode != "img2img":
        current_model_path = model_path

        if is_colab:
            pipe = StableDiffusionImg2ImgPipeline.from_pretrained(
                current_model_path, torch_dtype=torch.float16, scheduler=scheduler)
        else:
            pipe.to("cpu")
            pipe = current_model.pipe_i2i

        if torch.cuda.is_available():
            pipe = pipe.to("cuda")
        last_mode = "img2img"

    prompt = current_model.prefix + prompt
    ratio = min(height / img.height, width / img.width)
    img = img.resize(
        (int(img.width * ratio), int(img.height * ratio)), Image.LANCZOS)
    result = pipe(
        prompt,
        negative_prompt=neg_prompt,
        # num_images_per_prompt=n_images,
        init_image=img,
        num_inference_steps=int(steps),
        strength=strength,
        guidance_scale=guidance,
        width=width,
        height=height,
        generator=generator)

    return replace_nsfw_images(result)


def replace_nsfw_images(results):
    for i in range(len(results.images)):
        if results.nsfw_content_detected[i]:
            results.images[i] = Image.open("nsfw.png")
    return results.images[0]


css = """.playground-diffusion-div div{display:inline-flex;align-items:center;gap:.8rem;font-size:1.75rem}.playground-diffusion-div div h1{font-weight:900;margin-bottom:7px}.playground-diffusion-div p{margin-bottom:10px;font-size:94%}a{text-decoration:underline}.tabs{margin-top:0;margin-bottom:0}#gallery{min-height:20rem}
"""
with gr.Blocks(css=css) as demo:
    gr.HTML(
        f"""
            <div class="playground-diffusion-div">
              <div>
                <h1>Playground Diffusion</h1>
              </div>
              <p>
               Demo for multiple fine-tuned Stable Diffusion models, trained on different styles: <br>
               <a href="https://huggingface.co/riccardogiorato/avatar-diffusion">Avatar</a>,<br/>
               <a href="https://huggingface.co/riccardogiorato/beeple-diffusion">Beeple</a>,<br/>
               <a href="https://huggingface.co/s3nh/beksinski-style-stable-diffusion">Beksinski</a>,<br/>
               Diffusers 🧨 SD model hosted on HuggingFace 🤗.
              </p>
               Running on <b>{device}</b>{(" in a <b>Google Colab</b>." if is_colab else "")}
              </p>
            </div>
        """
    )
    with gr.Row():

        with gr.Column(scale=55):
            with gr.Group():
                model_name = gr.Dropdown(label="Model", choices=[
                                         m.name for m in models], value=current_model.name)

                with gr.Row():
                    prompt = gr.Textbox(label="Prompt", show_label=False, max_lines=2,
                                        placeholder="Enter prompt. Style applied automatically").style(container=False)
                    generate = gr.Button(value="Generate").style(
                        rounded=(False, True, True, False))

                image_out = gr.Image(height=512)
                # gallery = gr.Gallery(
                #     label="Generated images", show_label=False, elem_id="gallery"
                # ).style(grid=[1], height="auto")

        with gr.Column(scale=45):
            with gr.Tab("Options"):
                with gr.Group():
                    neg_prompt = gr.Textbox(
                        label="Negative prompt", placeholder="What to exclude from the image")

                    # n_images = gr.Slider(label="Images", value=1, minimum=1, maximum=4, step=1)

                    with gr.Row():
                        guidance = gr.Slider(
                            label="Guidance scale", value=7.5, maximum=15)
                        steps = gr.Slider(
                            label="Steps", value=25, minimum=2, maximum=75, step=1)

                    with gr.Row():
                        width = gr.Slider(
                            label="Width", value=512, minimum=64, maximum=1024, step=8)
                        height = gr.Slider(
                            label="Height", value=512, minimum=64, maximum=1024, step=8)

                    seed = gr.Slider(
                        0, 2147483647, label='Seed (0 = random)', value=0, step=1)

            with gr.Tab("Image to image"):
                with gr.Group():
                    image = gr.Image(label="Image", height=256,
                                     tool="editor", type="pil")
                    strength = gr.Slider(
                        label="Transformation strength", minimum=0, maximum=1, step=0.01, value=0.5)

    inputs = [model_name, prompt, guidance, steps,
              width, height, seed, image, strength, neg_prompt]
    prompt.submit(inference, inputs=inputs, outputs=image_out)
    generate.click(inference, inputs=inputs, outputs=image_out)

    ex = gr.Examples([
        [models[0].name, "Neon techno-magic robot with spear pierces an ancient beast, hyperrealism, no blur, 4k resolution, ultra detailed", 7.5, 50],
        [models[0].name, "halfturn portrait of a big crystal face of a beautiful abstract ancient Egyptian elderly shaman woman, made of iridescent golden crystals, half - turn, bottom view, ominous, intricate, studio, art by anthony macbain and greg rutkowski and alphonse mucha, concept art, 4k, sharp focus", 7.5, 25],
    ], [model_name, prompt, guidance, steps, seed], image_out, inference, cache_examples=False)

    gr.HTML("""
      <p>Models by <a href="https://huggingface.co/riccardogiorato">@riccardogiorato</a><br></p>
    """)

if not is_colab:
    demo.queue(concurrency_count=1)
demo.launch(debug=is_colab, share=is_colab)