import gradio as gr
import torch
import torch.nn as nn
import numpy as np

class Generator(nn.Module):
    def __init__(self):
        super(Generator, self).__init__()
        self.main = nn.Sequential(
            nn.ConvTranspose2d(128, 64 * 8, 4, 1, 0, bias=False),
            nn.BatchNorm2d(64 * 8),
            nn.ReLU(True),

            nn.ConvTranspose2d(64 * 8, 64 * 4, 4, 2, 1, bias=False),
            nn.BatchNorm2d(64 * 4),
            nn.ReLU(True),

            nn.ConvTranspose2d(64 * 4, 64 * 2, 4, 2, 1, bias=False),
            nn.BatchNorm2d(64 * 2),
            nn.ReLU(True),

            nn.ConvTranspose2d(64 * 2, 64, 4, 2, 1, bias=False),
            nn.BatchNorm2d(64),
            nn.ReLU(True),

            nn.ConvTranspose2d(64, 3, 4, 2, 1, bias=False),
            nn.Tanh()
        )

    def forward(self, input):
        return self.main(input)

netG = Generator()

device = "cuda" if torch.cuda.is_available() else "cpu"
model_file = "model.pth"
netG.load_state_dict(torch.load(model_file, map_location=device))
netG.eval()

def generate_image():
    with torch.no_grad():
        noise = torch.randn(1, 128, 1, 1)
        fake_image = netG(noise)

    generated_image = fake_image.squeeze().cpu().numpy()
    generated_image = np.transpose(generated_image, (1, 2, 0))
    generated_image = (generated_image + 1) / 2.0 
    generated_image = (generated_image * 255).astype(np.uint8)

    return generated_image

title = "DCGAN Image Generator 🖌️🎨"
description = "Generate non-existing images using DCGAN."
content = """
## How to Use 🎨

To generate an image, follow these steps:

1. Click \"Generate\" button to generate a image!
2. Once the image is generated, you can save it or share it to the community!
"""
    
iface = gr.Interface(
    fn=generate_image,
    inputs=None,
    outputs="image",
    title=title,
    description=description,
    article=content,
    api_name="generate"
)

iface.queue()
iface.launch()