import torch
from diffusers import StableDiffusion3Pipeline
import gradio as gr
import os
import spaces

HF_TOKEN = os.environ.get('HF_TOKEN')

if torch.cuda.is_available():
    device = "cuda"
    print("Using GPU")
else:
    device = "cpu"
    print("Using CPU")

# Initialize the pipeline and download the model
pipe = StableDiffusion3Pipeline.from_pretrained("stabilityai/stable-diffusion-3-medium", torch_dtype=torch.float16, use_auth_token=HF_TOKEN)
pipe.to(device)

# Define the image generation function
@spaces.GPU(duration=60)
def generate_image(prompt):
    image = pipe(
        prompt=prompt,
        negative_prompt=negative_prompt,
        num_inference_steps=num_inference_steps,
        height=height,
        width=width,
        guidance_scale=guidance_scale,
    ).images[0]
    return image

# Create the Gradio interface

prompt = gr.Textbox(label="Prompt", info="Describe the image you want", placeholder="A cat...")

negative_prompt = gr.Textbox(label="Negative Prompt", info="Describe what you don't want in the image", placeholder="Ugly, bad anatomy...")

num_inference_steps = gr.Number(label="Number of Inference Steps", precision=0, value=25)

height = gr.Number(label="Number of Inference Steps", precision=0, value=1024)

width = gr.Number(label="Number of Inference Steps", precision=0, value=1024)

guidance_scale = gr.Number(minimum=0.1, value=7.5, label="Guidance Scale", info="The number of denoising steps of the image. More denoising steps usually lead to a higher quality image at the cost of slower inference")

interface = gr.Interface(
    fn=generate_image,
    inputs=[prompt, negative_prompt, num_inference_steps, height, width, guidance_scale]
    outputs="image",
    title="Stable Diffusion 3 Medium",
    description="Made by [Nick088](https://linktr.ee/Nick088"
)

# Launch the interface
interface.launch(share=False)