app.py

import spaces
import torch
import gradio as gr
import os
from diffusers import FluxPipeline

# Initialize model outside the function
device = "cuda" if torch.cuda.is_available() else "cpu"
dtype = torch.bfloat16
file_url = "https://huggingface.co/lodestones/Chroma/blob/main/chroma-unlocked-v31.safetensors"
huggingface_token = os.getenv("HUGGINGFACE_TOKEN")

# Load the pipeline with proper configuration
flux_pipeline = FluxPipeline.from_single_file(
    file_url,
    torch_dtype=dtype,
    token=huggingface_token,
    use_safetensors=True,
    local_files_only=False,
    config_file="model_index.json"
)
flux_pipeline.to(device)

@spaces.GPU()
def generate_image(prompt, negative_prompt="", num_inference_steps=30, guidance_scale=7.5):
    # Generate image
    image = flux_pipeline(
        prompt=prompt,
        negative_prompt=negative_prompt,
        num_inference_steps=num_inference_steps,
        guidance_scale=guidance_scale
    ).images[0]
    
    return image

# Create Gradio interface
iface = gr.Interface(
    fn=generate_image,
    inputs=[
        gr.Textbox(label="Prompt", placeholder="Enter your prompt here..."),
        gr.Textbox(label="Negative Prompt", placeholder="Enter negative prompt here...", value=""),
        gr.Slider(minimum=1, maximum=100, value=30, step=1, label="Number of Inference Steps"),
        gr.Slider(minimum=1.0, maximum=20.0, value=7.5, step=0.1, label="Guidance Scale")
    ],
    outputs=gr.Image(label="Generated Image"),
    title="Chroma Image Generator",
    description="Generate images using the Chroma model",
    examples=[
        ["A beautiful sunset over mountains, photorealistic, 8k", "blurry, low quality, distorted", 30, 7.5],
        ["A futuristic cityscape at night, neon lights, cyberpunk style", "ugly, deformed, low resolution", 30, 7.5]
    ]
)

if __name__ == "__main__":
    iface.launch()