Update app.py

app.py

@@ -1,220 +1,29 @@
-import sys
-import subprocess
-import importlib.util
-import os
-
-# List of required packages.
-required_packages = {
-    "gradio": "gradio",
-    "numpy": "numpy",
-    "torch": "torch",
-    "diffusers": "diffusers",
-    "PIL": "pillow",
-    "spaces": "spaces",  # if this is a package available in your environment
-    "transformers": "transformers",
-}
-
-def install_package(package_name):
-    subprocess.check_call([sys.executable, "-m", "pip", "install", package_name])
-
-# Auto-install any missing packages.
-for mod, pkg in required_packages.items():
-    if importlib.util.find_spec(mod) is None:
-        print(f"Module {mod} not found, installing {pkg}...")
-        install_package(pkg)
-
-# Explicitly import transformers to ensure it's loaded.
-import transformers
-
-import random
 import gradio as gr
-import numpy as np
-import spaces
+from diffusers import StableDiffusionImg2ImgPipeline
 import torch
-from diffusers import DiffusionPipeline
 from PIL import Image
 
-device = "cuda" if torch.cuda.is_available() else "cpu"
-
-# Model identifiers.
-repo_id = "black-forest-labs/FLUX.1-dev"
-adapter_id = "alvarobartt/ghibli-characters-flux-lora"
-
-# Retrieve HF token from environment (if required to access gated repositories).
-hf_token = os.environ.get("HF_TOKEN", None)
-
-# Load the base model from the repository.
-pipeline = DiffusionPipeline.from_pretrained(
-    repo_id,
-    torch_dtype=torch.bfloat16,
-    use_auth_token=hf_token  # Only needed if the repo is gated.
+# Load the fine-tuned Studio Ghibli style model
+model_id = "nitrosocke/Ghibli-Diffusion"  # This model is fine-tuned to mimic the Studio Ghibli aesthetic
+pipe = StableDiffusionImg2ImgPipeline.from_pretrained(model_id, torch_dtype=torch.float16)
+pipe.to("cuda")  # Ensure you have GPU support in your Space; if not, remove or change accordingly
+
+def transform_image(input_image: Image.Image) -> Image.Image:
+    # Resize input to a standard size (e.g., 512x512)
+    input_image = input_image.resize((512, 512))
+    prompt = "ghibli style"
+    # 'strength' controls how much the input image is modified (0.0 to 1.0)
+    output = pipe(prompt=prompt, image=input_image, strength=0.75, guidance_scale=7.5)
+    return output.images[0]
+
+# Create a Gradio interface: input is an image and output is the transformed image.
+demo = gr.Interface(
+    fn=transform_image,
+    inputs=gr.Image(type="pil", label="Upload your portrait/photo"),
+    outputs=gr.Image(type="pil", label="Studio Ghibli Style Output"),
+    title="Studio Ghibli Style Converter",
+    description="Upload a portrait or photo and transform it into a Studio Ghibli-style image using a fine-tuned Stable Diffusion model."
 )
-pipeline.load_lora_weights(adapter_id)
-pipeline = pipeline.to(device)
-
-MAX_SEED = np.iinfo(np.int32).max
-MAX_IMAGE_SIZE = 1024
-
-@spaces.GPU(duration=80)
-def inference(
-    prompt: str,
-    seed: int,
-    randomize_seed: bool,
-    width: int,
-    height: int,
-    guidance_scale: float,
-    num_inference_steps: int,
-    lora_scale: float,
-    progress: gr.Progress = gr.Progress(track_tqdm=True),
-):
-    if randomize_seed:
-        seed = random.randint(0, MAX_SEED)
-    generator = torch.Generator(device=device).manual_seed(seed)
-    
-    image = pipeline(
-        prompt=prompt,
-        guidance_scale=guidance_scale,
-        num_inference_steps=num_inference_steps,
-        width=width,
-        height=height,
-        generator=generator,
-        joint_attention_kwargs={"scale": lora_scale},
-    ).images[0]
-
-    return image, seed
-
-examples = [
-    (
-        "Ghibli style futuristic stormtrooper with glossy white armor and a sleek helmet,"
-        " standing heroically on a lush alien planet, vibrant flowers blooming around, soft"
-        " sunlight illuminating the scene, a gentle breeze rustling the leaves"
-    ),
-]
-
-css = """
-#col-container {
-    margin: 0 auto;
-    max-width: 640px;
-}
-"""
-
-with gr.Blocks(css=css) as demo:
-    with gr.Column(elem_id="col-container"):
-        gr.Markdown("# FLUX.1 Studio Ghibli LoRA")
-        gr.Markdown(
-            "[alvarobartt/ghibli-characters-flux-lora](https://huggingface.co/alvarobartt/ghibli-characters-flux-lora)"
-            " is a LoRA fine-tune of [FLUX.1-dev](https://huggingface.co/black-forest-labs/FLUX.1-dev)"
-            " with [alvarobartt/ghibli-characters](https://huggingface.co/datasets/alvarobartt/ghibli-characters)."
-        )
-
-        with gr.Accordion("How to generate nice prompts?", open=False):
-            gr.Markdown(
-                "What worked best for me to generate high-quality prompts of well-known characters,"
-                " was to prompt either [Claude 3 Haiku](https://claude.ai), [GPT4-o](https://chatgpt.com/),"
-                " or [Perplexity](https://www.perplexity.ai/) with:\n\nYou are an"
-                " expert prompt writer for diffusion text to image models, and you've been provided"
-                " the following prompt template:\n\n\"Ghibli style [character description] with"
-                " [distinctive features], [action or pose], [environment or background],"
-                " [lighting or atmosphere], [additional details].\"\n\nCould you create a prompt"
-                " to generate [CHARACTER NAME] as a Studio Ghibli character following that template? "
-                "[MORE DETAILS IF NEEDED]\n"
-            )
-
-        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=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=768,
-                )
-
-            with gr.Row():
-                guidance_scale = gr.Slider(
-                    label="Guidance scale",
-                    minimum=0.0,
-                    maximum=10.0,
-                    step=0.1,
-                    value=3.5,
-                )
-
-                num_inference_steps = gr.Slider(
-                    label="Number of inference steps",
-                    minimum=1,
-                    maximum=50,
-                    step=1,
-                    value=30,
-                )
-
-                lora_scale = gr.Slider(
-                    label="LoRA scale",
-                    minimum=0.0,
-                    maximum=1.0,
-                    step=0.1,
-                    value=1.0,
-                )
-
-        gr.Examples(
-            examples=examples,
-            fn=lambda x: (Image.open("./example.jpg"), 42),
-            inputs=[prompt],
-            outputs=[result, seed],
-            run_on_click=True,
-        )
-
-        gr.Markdown(
-            "### Disclaimer\n\n"
-            "License is non-commercial for both FLUX.1-dev and the Studio Ghibli dataset; "
-            "but free to use for personal and non-commercial purposes."
-        )
-
-    gr.on(
-        triggers=[run_button.click, prompt.submit],
-        fn=inference,
-        inputs=[
-            prompt,
-            seed,
-            randomize_seed,
-            width,
-            height,
-            guidance_scale,
-            num_inference_steps,
-            lora_scale,
-        ],
-        outputs=[result, seed],
-    )
 
-demo.queue()
-demo.launch()
+if __name__ == "__main__":
+    demo.launch()