app.py

import gradio as gr
import torch
from diffusers import DiffusionPipeline
import os

# Ensure GPU hardware is selected in Space settings!
print("Loading diffusion pipeline...")
pipe = DiffusionPipeline.from_pretrained(
    "John6666/the-araminta-fv1-sdxl",
    torch_dtype=torch.float16,
    use_safetensors=True,
    token=os.environ.get("HF_TOKEN") # Use token from secrets or OAuth
)
if torch.cuda.is_available():
    pipe = pipe.to("cuda")
else:
    print("WARNING: No CUDA GPU available, using CPU.")

def generate_image(prompt, negative_prompt="", guidance_scale=7.5, num_steps=30):
    print(f"Generating image for prompt: {prompt}")
    try:
        image = pipe(
            prompt=prompt,
            negative_prompt=negative_prompt,
            guidance_scale=guidance_scale,
            num_inference_steps=int(num_steps) # Ensure steps is int
        ).images[0]
        print("Image generation successful.")
        return image
    except Exception as e:
        print(f"Error during image generation: {e}")
        raise gr.Error(f"Failed to generate image: {e}")

with gr.Blocks() as demo:
    gr.Markdown("# The Araminta FV1 SDXL")
    with gr.Row():
        with gr.Column():
            prompt_input = gr.Textbox(label="Prompt")
            neg_prompt_input = gr.Textbox(label="Negative Prompt", value="")
            gs_input = gr.Slider(label="Guidance Scale", minimum=1, maximum=20, step=0.5, value=7.5)
            steps_input = gr.Slider(label="Inference Steps", minimum=10, maximum=100, step=1, value=30)
            submit_btn = gr.Button("Generate")
        with gr.Column():
            output_image = gr.Image(label="Generated Image")

    submit_btn.click(
        fn=generate_image,
        inputs=[prompt_input, neg_prompt_input, gs_input, steps_input],
        outputs=output_image
    )

print("Launching Gradio interface...")
demo.launch()