app.py

from diffusers import DiffusionPipeline
import torch
import os

try:
    import intel_extension_for_pytorch as ipex
except:
    pass

from PIL import Image
import numpy as np
import gradio as gr
import psutil
import time

SAFETY_CHECKER = os.environ.get("SAFETY_CHECKER", None)
TORCH_COMPILE = os.environ.get("TORCH_COMPILE", None)
HF_TOKEN = os.environ.get("HF_TOKEN", None)
# check if MPS is available OSX only M1/M2/M3 chips
mps_available = hasattr(torch.backends, "mps") and torch.backends.mps.is_available()
xpu_available = hasattr(torch, "xpu") and torch.xpu.is_available()
device = torch.device(
    "cuda" if torch.cuda.is_available() else "xpu" if xpu_available else "cpu"
)
torch_device = device
torch_dtype = torch.float16

print(f"SAFETY_CHECKER: {SAFETY_CHECKER}")
print(f"TORCH_COMPILE: {TORCH_COMPILE}")
print(f"device: {device}")

if mps_available:
    device = torch.device("mps")
    torch_device = "cpu"
    torch_dtype = torch.float32

if SAFETY_CHECKER == "True":
    pipe = DiffusionPipeline.from_pretrained(
        "stabilityai/sdxl-turbo",
        torch_dtype=torch_dtype,
        variant="fp16" if torch_dtype == torch.float16 else "fp32")
else:
    pipe = DiffusionPipeline.from_pretrained(
        "stabilityai/sdxl-turbo",
        safety_checker=None,
        torch_dtype=torch_dtype,
        variant="fp16" if torch_dtype == torch.float16 else "fp32",
    )


pipe.to(device=torch_device, dtype=torch_dtype).to(device)
pipe.unet.to(memory_format=torch.channels_last)
pipe.set_progress_bar_config(disable=True)


def predict(prompt, steps, seed=1231231):
    generator = torch.manual_seed(seed)
    last_time = time.time()
    results = pipe(
        prompt=prompt,
        generator=generator,
        num_inference_steps=steps,
        guidance_scale=0.0,
        width=512,
        height=512,
        # original_inference_steps=params.lcm_steps,
        output_type="pil",
    )
    print(f"Pipe took {time.time() - last_time} seconds")
    nsfw_content_detected = (
        results.nsfw_content_detected[0]
        if "nsfw_content_detected" in results
        else False
    )
    if nsfw_content_detected:
        gr.Warning("NSFW content detected.")
        return Image.new("RGB", (512, 512))
    return results.images[0]


css = """
#container{
    margin: 0 auto;
    max-width: 40rem;
}
#intro{
    max-width: 100%;
    text-align: center;
    margin: 0 auto;
}
"""
with gr.Blocks(css=css) as demo:
    with gr.Column(elem_id="container"):
        gr.Markdown(
            """# SDXL Turbo - Text To Image
            ## Unofficial Demo
            SDXL Turbo model can generate high quality images in a single pass read more on [stability.ai post](https://stability.ai/news/stability-ai-sdxl-turbo).  
            **Model**: https://huggingface.co/stabilityai/sdxl-turbo
            """,
            elem_id="intro",
        )
        with gr.Row():
            with gr.Row():
                prompt = gr.Textbox(
                    placeholder="Insert your prompt here:", scale=5, container=False
                )
                generate_bt = gr.Button("Generate", scale=1)

        image = gr.Image(type="filepath")
        with gr.Accordion("Advanced options", open=False):
            steps = gr.Slider(label="Steps", value=2, minimum=1, maximum=10, step=1)
            seed = gr.Slider(
                randomize=True, minimum=0, maximum=12013012031030, label="Seed", step=1
            )
        with gr.Accordion("Run with diffusers"):
            gr.Markdown(
                """## Running SDXL Turbo with `diffusers`
            ```bash
            pip install diffusers==0.23.1
            ```
            ```py
            from diffusers import DiffusionPipeline

            pipe = DiffusionPipeline.from_pretrained(
                "stabilityai/sdxl-turbo", variant="fp16", torch_dtype=torch.float16
            ).to("cuda")
            results = pipe(
                prompt="A cinematic shot of a baby racoon wearing an intricate italian priest robe",
                num_inference_steps=1,
                guidance_scale=0.0,
            )
            imga = results.images[0]
            imga.save("image.png")
            ```
            """
            )

        inputs = [prompt, steps, seed]
        generate_bt.click(fn=predict, inputs=inputs, outputs=image, show_progress=False)
        prompt.input(fn=predict, inputs=inputs, outputs=image, show_progress=False)
        steps.change(fn=predict, inputs=inputs, outputs=image, show_progress=False)
        seed.change(fn=predict, inputs=inputs, outputs=image, show_progress=False)

demo.queue()
demo.launch()