app.py

import gradio as gr
import numpy as np
import random
import spaces
import torch
from PIL import Image
from torchvision import transforms
from diffusers import DiffusionPipeline

# Define constants
dtype = torch.bfloat16
device = "cuda" if torch.cuda.is_available() else "cpu"
MAX_SEED = np.iinfo(np.int32).max
MAX_IMAGE_SIZE = 2048

# Load the diffusion pipeline
pipe = DiffusionPipeline.from_pretrained("black-forest-labs/FLUX.1-schnell", torch_dtype=dtype).to(device)

def preprocess_image(image, image_size):
    print(f"Preprocessing image to size: {image_size}x{image_size}")
    preprocess = transforms.Compose([
        transforms.Resize((image_size, image_size)),  # Use model-specific size
        transforms.ToTensor(),
        transforms.Normalize([0.5], [0.5])  # Ensure this matches the VAE's training normalization
    ])
    image = preprocess(image).unsqueeze(0).to(device, dtype=dtype)
    print(f"Image shape after preprocessing: {image.shape}")
    return image

def encode_image(image, vae):
    print("Encoding image using the VAE")
    with torch.no_grad():
        latents = vae.encode(image).latent_dist.sample() * 0.18215
    print(f"Latents shape after encoding: {latents.shape}")
    return latents

@spaces.GPU()
def infer(prompt, init_image=None, seed=42, randomize_seed=False, width=1024, height=1024, num_inference_steps=4, progress=gr.Progress(track_tqdm=True)):
    print(f"Inference started with prompt: {prompt}")
    if randomize_seed:
        seed = random.randint(0, MAX_SEED)
    print(f"Using seed: {seed}")
    generator = torch.Generator().manual_seed(seed)

    if init_image is None:
        print("No initial image provided, processing text2img")
        # Process text2img
        try:
            print("Calling the diffusion pipeline without latents")
            result = pipe(
                prompt=prompt,
                height=height,
                width=width,
                num_inference_steps=num_inference_steps,
                generator=generator,
                guidance_scale=0.0
            )
            image = result.images[0]
            latents = result.latents
            
            # Log the latent shapes from text2img process
            print(f"Latents shape from text2img: {latents.shape}")
        except Exception as e:
            print(f"Pipeline call failed with error: {e}")
            raise

    else:
        print("Initial image provided, processing img2img")
        vae_image_size = pipe.vae.config.sample_size
        print(f"Expected VAE image size: {vae_image_size}")
        init_image = init_image.convert("RGB")
        init_image = preprocess_image(init_image, vae_image_size)
        latents = encode_image(init_image, pipe.vae)

        # Interpolating latents
        print(f"Interpolating latents to size: {(height // 8, width // 8)}")
        latents = torch.nn.functional.interpolate(latents, size=(height // 8, width // 8))
        print(f"Latents shape after interpolation: {latents.shape}")
        
        # Convert latent channels to 64 as expected by the transformer
        latent_channels = pipe.vae.config.latent_channels
        print(f"Expected latent channels: 64, current latent channels: {latent_channels}")
        if latent_channels != 64:
            print(f"Converting latent channels from {latent_channels} to 64")
            conv = torch.nn.Conv2d(latent_channels, 64, kernel_size=1).to(device, dtype=dtype)
            latents = conv(latents)
            print(f"Latents shape after channel conversion: {latents.shape}")

        # Debugging input shape before calling transformer
        print(f"Latents shape before reshaping for transformer: {latents.shape}")

        # Reshape latents to match the transformer's input expectations
        latents = latents.permute(0, 2, 3, 1).contiguous().view(-1, 64)  # Assuming the transformer expects (batch, sequence, feature)
        print(f"Latents shape after reshaping for transformer: {latents.shape}")

        # Adding extra debug to understand what transformer expects
        try:
            print("Calling the transformer with latents")
            # Dummy call to transformer to understand the shape requirement
            _ = pipe.transformer(latents)
            print("Transformer call succeeded")
        except Exception as e:
            print(f"Transformer call failed with error: {e}")
            raise

        print("Calling the diffusion pipeline with latents")
        image = pipe(
            prompt=prompt,
            height=height,
            width=width,
            num_inference_steps=num_inference_steps,
            generator=generator,
            guidance_scale=0.0,
            latents=latents
        ).images[0]
    
    print("Inference complete")
    return image, seed






# Define example prompts
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 styling for the Japanese-inspired interface
css = """
body {
    background-color: #fff;
    font-family: 'Noto Sans JP', sans-serif;
    color: #333;
}
#col-container {
    margin: 0 auto;
    max-width: 520px;
    border: 2px solid #000;
    padding: 20px;
    background-color: #f7f7f7;
    border-radius: 10px;
}
.gr-button {
    background-color: #e60012;
    color: #fff;
    border: 2px solid #000;
}
.gr-button:hover {
    background-color: #c20010;
}
.gr-slider, .gr-checkbox, .gr-textbox {
    border: 2px solid #000;
}
.gr-accordion {
    border: 2px solid #000;
    background-color: #fff;
}
.gr-image {
    border: 2px solid #000;
}
"""

# Create the Gradio interface
with gr.Blocks(css=css) as demo:
    
    with gr.Column(elem_id="col-container"):
        gr.Markdown("""
        # FLUX.1 [schnell]
        12B param rectified flow transformer distilled from [FLUX.1 [pro]](https://blackforestlabs.ai/) for 4 step generation
        [[blog](https://blackforestlabs.ai/announcing-black-forest-labs/)] [[model](https://huggingface.co/black-forest-labs/FLUX.1-schnell)]
        """)

        with gr.Row():
            prompt = gr.Textbox(
                label="Prompt",
                show_label=False,
                max_lines=1,
                placeholder="Enter your prompt",
                container=False,
            )
            run_button = gr.Button("Run", scale=0)

        with gr.Row():
            init_image = gr.Image(label="Initial Image (optional)", type="pil")
            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=42,
            )
            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():
                num_inference_steps = gr.Slider(
                    label="Number of inference steps",
                    minimum=1,
                    maximum=50,
                    step=1,
                    value=4,
                )

        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, init_image, seed, randomize_seed, width, height, num_inference_steps],
        outputs=[result, seed]
    )

demo.launch()