app.py

import os
import gradio as gr
import numpy as np
import random
import spaces
import torch
#from diffusers import DiffusionPipeline
from diffusers import AutoPipelineForImage2Image
from huggingface_hub import InferenceClient

dtype = torch.bfloat16
device = "cuda" 
#if torch.cuda.is_available() else "cpu"

#pipe = DiffusionPipeline.from_pretrained("black-forest-labs/FLUX.1-dev", torch_dtype=torch.bfloat16).to(device)
sdxl = InferenceClient(model="stabilityai/stable-diffusion-xl-base-1.0", token=os.environ['HF_TOKEN'])
print('sdxl loaded')

"kandinsky-community/kandinsky-2-2-decoder"
#pipeline2Image = DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-xl-refiner-1.0", torch_dtypes=torch.bfloat16).to(device)
#pipeline2Image = DiffusionPipeline.from_pretrained("kandinsky-community/kandinsky-2-2-decoder", torch_dtypes=torch.bfloat16).to(device)
pipeline2Image = AutoPipelineForImage2Image.from_pretrained("kandinsky-community/kandinsky-2-2-decoder", torch_dtype=dtype)
pipeline2Image.enable_model_cpu_offload()
print("pipeline 2 image loaded")

MAX_SEED = np.iinfo(np.int32).max
MAX_IMAGE_SIZE = 2048

# (duration=190)
#@spaces.GPU
def infer(prompt, seed=42, randomize_seed=False, width=1024, height=1024, guidance_scale=5.0, num_inference_steps=28, progress=gr.Progress(track_tqdm=True)):
    if randomize_seed:
        seed = random.randint(0, MAX_SEED)
#    generator = torch.Generator().manual_seed(seed)
#    image = pipe(
#        prompt=prompt, 
#        width=width,
#        height=height,
#        num_inference_steps=num_inference_steps, 
#        generator=generator,
#        guidance_scale=guidance_scale
#    ).images[0] 
    image = sdxl.text_to_image(
        prompt,
        guidance_scale=guidance_scale, num_inference_steps=num_inference_steps, seed=seed,width=width, height=height
    )
    
    return image, seed

examples = [
    "a tiny astronaut hatching from an egg on the moon",
    "a cat holding a sign that says hello world",
    "an anime illustration of a wiener schnitzel",
]

css="""
#col-container {
    margin: 0 auto;
    max-width: 520px;
}
"""

with gr.Blocks(css=css) as demo:
    
    with gr.Column(elem_id="col-container"):
        gr.Markdown(f"""# FLUX.1 [dev]
12B param rectified flow transformer guidance-distilled from [FLUX.1 [pro]](https://blackforestlabs.ai/)  
[[non-commercial license](https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md)] [[blog](https://blackforestlabs.ai/announcing-black-forest-labs/)] [[model](https://huggingface.co/black-forest-labs/FLUX.1-dev)]
        """)
        
        with gr.Row():
            
            prompt = gr.Text(
                label="Prompt",
                show_label=False,
                max_lines=1,
                placeholder="Enter your prompt",
                container=False,
            )
            
            run_button = gr.Button("Run", scale=0)

        result = gr.Image(label="Result", show_label=False)
        
        with gr.Accordion("Advanced Settings", open=False):
            
            seed = gr.Slider(
                label="Seed",
                minimum=0,
                maximum=MAX_SEED,
                step=1,
                value=0,
            )
            
            randomize_seed = gr.Checkbox(label="Randomize seed", value=True)
            
            with gr.Row():
                
                width = gr.Slider(
                    label="Width",
                    minimum=256,
                    maximum=MAX_IMAGE_SIZE,
                    step=32,
                    value=1024,
                )
                
                height = gr.Slider(
                    label="Height",
                    minimum=256,
                    maximum=MAX_IMAGE_SIZE,
                    step=32,
                    value=1024,
                )
            
            with gr.Row():

                guidance_scale = gr.Slider(
                    label="Guidance Scale",
                    minimum=1,
                    maximum=15,
                    step=0.1,
                    value=3.5,
                )
  
                num_inference_steps = gr.Slider(
                    label="Number of inference steps",
                    minimum=1,
                    maximum=50,
                    step=1,
                    value=28,
                )
        
        gr.Examples(
            examples=examples,
            fn=infer,
            inputs=[prompt],
            outputs=[result, seed],
            cache_examples="lazy"
        )

    gr.on(
        triggers=[run_button.click, prompt.submit],
        fn=infer,
        inputs=[prompt, seed, randomize_seed, width, height, guidance_scale, num_inference_steps],
        outputs=[result, seed]
    )

    # Adding image input options at the bottom
    gr.Markdown("## Upload or select an additional image")
    
    with gr.Row():
        uploaded_image = gr.Image(label="Upload Image", type="pil")
        image_url = gr.Textbox(label="Image URL", placeholder="Enter image URL")
        use_generated_image = gr.Button("Use Generated Image")

    with gr.Accordion("Advanced Settings", open=False):
            
        seed2 = gr.Slider(
                label="Seed",
                minimum=0,
                maximum=MAX_SEED,
                step=1,
                value=0,
            )
            
        randomize_seed2 = gr.Checkbox(label="Randomize seed", value=True)
            
        with gr.Row():
                
            width2 = gr.Slider(
                    label="Width",
                    minimum=256,
                    maximum=MAX_IMAGE_SIZE,
                    step=32,
                    value=1024,
                )
                
            height2 = gr.Slider(
                    label="Height",
                    minimum=256,
                    maximum=MAX_IMAGE_SIZE,
                    step=32,
                    value=1024,
                )
            
        with gr.Row():

            strength2 = gr.Slider(
                    label="Strength",
                    minimum=.1,
                    maximum=1,
                    step=0.1,
                    value=.5,
                )
            guidance_scale2 = gr.Slider(
                    label="Guidance Scale",
                    minimum=1,
                    maximum=15,
                    step=0.1,
                    value=3.5,
                )
  
            num_inference_steps2 = gr.Slider(
                    label="Number of inference steps",
                    minimum=1,
                    maximum=50,
                    step=1,
                    value=28,
                )

    prompt2 = gr.Text(
                label="Prompt",
                show_label=False,
                max_lines=1,
                placeholder="Enter your prompt",
                container=False,
            )
    run2_button = gr.Button("Run", scale=0)
    additional_image_output = gr.Image(label="Selected Image", show_label=False)

    def select_image(uploaded_image, image_url, use_generated=False):
        if use_generated:
            return result.value
        elif uploaded_image is not None:
            return uploaded_image
        elif image_url:
            try:
                img = gr.Image.load(image_url)
                return img
            except Exception as e:
                return f"Failed to load image from URL: {e}"
        return None

    def image2image(uploaded_image, image_url, use_generated=False):
        image = select_image(uploaded_image, image_url, use_generated=use_generated)
        #prompt = "one awesome dude"
        #generator = torch.Generator(device=device).manual_seed(1024)
        #image = pipeline2Image(prompt=prompt, image=image, strength=0.75, guidance_scale=7.5, generator=generator).images[0]
        return image
    
    use_generated_image.click(fn=lambda: image2image(None, None, True), inputs=[], outputs=additional_image_output)
    uploaded_image.change(fn=image2image, inputs=[uploaded_image, image_url, gr.State(False)], outputs=additional_image_output)
    image_url.submit(fn=image2image, inputs=[uploaded_image, image_url, gr.State(False)], outputs=additional_image_output)

    @spaces.GPU(duration=190)
    def infer2(prompt, image, seed=42, randomize_seed=False, width=1024, height=1024, strength=.5, guidance_scale=5.0, num_inference_steps=28):
        if randomize_seed:
            seed = random.randint(0, MAX_SEED)
        generator = torch.Generator(device=device).manual_seed(seed)
        image2 = pipeline2Image(prompt=prompt, image=image, strength=strength, guidance_scale=guidance_scale, num_inference_steps=num_inference_steps, generator=generator).images[0]
#    generator = torch.Generator().manual_seed(seed)
#    image = pipe(
#        prompt=prompt, 
#        width=width,
#        height=height,
#        num_inference_steps=num_inference_steps, 
#        generator=generator,
#        guidance_scale=guidance_scale
#    ).images[0] 
        return image2, seed
    final_image_output = gr.Image(label="Final Image", show_label=False)
    
    gr.on(
        triggers=[run2_button.click, prompt2.submit],
        fn=infer2,
        inputs=[prompt2, torch.from_numpy(additional_image_output), seed2, randomize_seed2, width2, height2, strength2, guidance_scale2, num_inference_steps2],
        outputs=[final_image_output, seed2]
    )

demo.launch()