Update app.py
Browse files
app.py
CHANGED
@@ -1,55 +1,72 @@
|
|
1 |
import gradio as gr
|
2 |
-
from
|
3 |
import torch
|
4 |
-
from
|
5 |
-
|
6 |
-
|
7 |
-
|
8 |
-
|
9 |
-
|
10 |
-
|
11 |
-
|
12 |
-
|
13 |
-
|
14 |
-
|
15 |
-
|
16 |
-
|
17 |
-
|
18 |
-
|
19 |
-
|
20 |
-
|
21 |
-
|
22 |
-
|
23 |
-
|
24 |
-
|
25 |
-
|
26 |
-
|
27 |
-
|
28 |
-
|
29 |
-
|
30 |
-
|
31 |
-
|
32 |
-
|
33 |
-
|
34 |
-
|
35 |
-
|
36 |
-
|
37 |
-
|
38 |
-
|
39 |
-
)
|
40 |
-
|
41 |
-
|
42 |
-
|
43 |
-
|
44 |
-
|
45 |
-
|
46 |
-
|
47 |
-
|
48 |
-
|
49 |
-
|
50 |
-
|
51 |
-
|
52 |
-
|
53 |
-
|
54 |
-
|
55 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
import gradio as gr
|
2 |
+
from huggingface_hub import hf_hub_download
|
3 |
import torch
|
4 |
+
from PIL import Image
|
5 |
+
from torchvision import transforms
|
6 |
+
from skimage.color import rgb2lab, lab2rgb
|
7 |
+
import numpy as np
|
8 |
+
import requests
|
9 |
+
from io import BytesIO
|
10 |
+
|
11 |
+
# Download the model from Hugging Face Hub
|
12 |
+
repo_id = "Hammad712/GAN-Colorization-Model"
|
13 |
+
model_filename = "generator.pt"
|
14 |
+
model_path = hf_hub_download(repo_id=repo_id, filename=model_filename)
|
15 |
+
|
16 |
+
# Define the generator model (same architecture as used during training)
|
17 |
+
from fastai.vision.learner import create_body
|
18 |
+
from torchvision.models import resnet34
|
19 |
+
from fastai.vision.models.unet import DynamicUnet
|
20 |
+
|
21 |
+
def build_generator(n_input=1, n_output=2, size=256):
|
22 |
+
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
23 |
+
backbone = create_body(resnet34(), pretrained=True, n_in=n_input, cut=-2)
|
24 |
+
G_net = DynamicUnet(backbone, n_output, (size, size)).to(device)
|
25 |
+
return G_net
|
26 |
+
|
27 |
+
# Initialize and load the model
|
28 |
+
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
29 |
+
G_net = build_generator(n_input=1, n_output=2, size=256)
|
30 |
+
G_net.load_state_dict(torch.load(model_path, map_location=device))
|
31 |
+
G_net.eval()
|
32 |
+
|
33 |
+
# Preprocessing function
|
34 |
+
def preprocess_image(img):
|
35 |
+
img = img.convert("RGB")
|
36 |
+
img = transforms.Resize((256, 256), Image.BICUBIC)(img)
|
37 |
+
img = np.array(img)
|
38 |
+
img_to_lab = rgb2lab(img).astype("float32")
|
39 |
+
img_to_lab = transforms.ToTensor()(img_to_lab)
|
40 |
+
L = img_to_lab[[0], ...] / 50. - 1.
|
41 |
+
return L.unsqueeze(0).to(device)
|
42 |
+
|
43 |
+
# Inference function
|
44 |
+
def colorize_image(img, model):
|
45 |
+
L = preprocess_image(img)
|
46 |
+
with torch.no_grad():
|
47 |
+
ab = model(L)
|
48 |
+
L = (L + 1.) * 50.
|
49 |
+
ab = ab * 110.
|
50 |
+
Lab = torch.cat([L, ab], dim=1).permute(0, 2, 3, 1).cpu().numpy()
|
51 |
+
rgb_imgs = []
|
52 |
+
for img in Lab:
|
53 |
+
img_rgb = lab2rgb(img)
|
54 |
+
rgb_imgs.append(img_rgb)
|
55 |
+
return np.stack(rgb_imgs, axis=0)[0] # Return the first (and only) image
|
56 |
+
|
57 |
+
# Gradio interface function
|
58 |
+
def colorization_function(image):
|
59 |
+
colorized_image = colorize_image(image, G_net)
|
60 |
+
return colorized_image
|
61 |
+
|
62 |
+
# Gradio Interface Setup
|
63 |
+
iface = gr.Interface(
|
64 |
+
fn=colorization_function,
|
65 |
+
inputs=gr.inputs.Image(type="pil", label="Upload Grayscale Image"),
|
66 |
+
outputs=gr.outputs.Image(type="numpy", label="Colorized Image"),
|
67 |
+
title="Image Colorization with GAN",
|
68 |
+
description="Upload a grayscale image, and the model will colorize it using AI."
|
69 |
+
)
|
70 |
+
|
71 |
+
# Launch the Gradio interface
|
72 |
+
iface.launch()
|