Spaces:
Running
Running
Update app.py
Browse files
app.py
CHANGED
@@ -1,71 +1,80 @@
|
|
1 |
import torch
|
|
|
2 |
import gradio as gr
|
3 |
import cv2
|
4 |
import numpy as np
|
5 |
|
6 |
-
#
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
7 |
MODEL_PATH = 'ColorizeVideo_gen.pth'
|
8 |
|
9 |
-
#
|
10 |
def load_model(model_path):
|
11 |
-
|
12 |
-
model
|
13 |
-
model.eval()
|
14 |
return model
|
15 |
|
16 |
-
#
|
17 |
def preprocess_frame(frame):
|
18 |
-
#
|
19 |
-
frame = cv2.resize(frame, (224, 224)) #
|
20 |
-
frame = frame / 255.0 #
|
21 |
-
input_tensor = torch.from_numpy(frame.astype(np.float32)).permute(2, 0, 1)
|
22 |
-
return input_tensor.unsqueeze(0)
|
23 |
|
24 |
-
#
|
25 |
def process_video(model, video_path):
|
26 |
cap = cv2.VideoCapture(video_path)
|
27 |
-
width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
|
28 |
-
height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
|
29 |
fourcc = cv2.VideoWriter_fourcc(*'mp4v')
|
30 |
output_path = "output_video.mp4"
|
31 |
-
out = cv2.VideoWriter(output_path, fourcc, 30.0, (
|
32 |
|
33 |
while cap.isOpened():
|
34 |
ret, frame = cap.read()
|
35 |
if not ret:
|
36 |
break
|
37 |
|
38 |
-
#
|
39 |
input_tensor = preprocess_frame(frame)
|
40 |
|
41 |
-
#
|
42 |
with torch.no_grad():
|
43 |
predictions = model(input_tensor)
|
44 |
|
45 |
-
#
|
46 |
output_frame = (predictions.squeeze().permute(1, 2, 0).numpy() * 255).astype(np.uint8)
|
47 |
-
output_frame = cv2.resize(output_frame, (frame.shape[1], frame.shape[0])) # Rétablir la taille originale
|
48 |
|
49 |
-
#
|
50 |
out.write(output_frame)
|
51 |
|
52 |
cap.release()
|
53 |
out.release()
|
54 |
return output_path
|
55 |
|
56 |
-
#
|
57 |
def colorize_video(video):
|
58 |
model = load_model(MODEL_PATH)
|
59 |
-
output_video_path = process_video(model, video.name)
|
60 |
return output_video_path
|
61 |
|
62 |
-
#
|
63 |
iface = gr.Interface(
|
64 |
fn=colorize_video,
|
65 |
-
inputs=gr.Video(label="
|
66 |
-
outputs=gr.Video(label="
|
67 |
-
title="
|
68 |
-
description="
|
69 |
)
|
70 |
|
71 |
if __name__ == '__main__':
|
|
|
1 |
import torch
|
2 |
+
import torch.nn as nn
|
3 |
import gradio as gr
|
4 |
import cv2
|
5 |
import numpy as np
|
6 |
|
7 |
+
# Define your model architecture
|
8 |
+
class YourModelArchitecture(nn.Module):
|
9 |
+
def __init__(self):
|
10 |
+
super(YourModelArchitecture, self).__init__()
|
11 |
+
# Define the layers of your model here
|
12 |
+
# Example: self.conv1 = nn.Conv2d(3, 16, kernel_size=3, stride=1, padding=1)
|
13 |
+
|
14 |
+
def forward(self, x):
|
15 |
+
# Define the forward pass logic
|
16 |
+
return x
|
17 |
+
|
18 |
+
# Path to the model weights
|
19 |
MODEL_PATH = 'ColorizeVideo_gen.pth'
|
20 |
|
21 |
+
# Load the model function
|
22 |
def load_model(model_path):
|
23 |
+
model = YourModelArchitecture() # Initialize the model architecture
|
24 |
+
model.load_state_dict(torch.load(model_path, map_location=torch.device('cpu'))) # Load the model weights
|
25 |
+
model.eval() # Set the model to evaluation mode
|
26 |
return model
|
27 |
|
28 |
+
# Preprocess the frame before passing it to the model
|
29 |
def preprocess_frame(frame):
|
30 |
+
# Resize and normalize the image
|
31 |
+
frame = cv2.resize(frame, (224, 224)) # Resize to model input size
|
32 |
+
frame = frame / 255.0 # Normalize the pixel values to [0, 1]
|
33 |
+
input_tensor = torch.from_numpy(frame.astype(np.float32)).permute(2, 0, 1) # Convert to tensor and change the dimension order
|
34 |
+
return input_tensor.unsqueeze(0) # Add batch dimension
|
35 |
|
36 |
+
# Process the video, frame by frame
|
37 |
def process_video(model, video_path):
|
38 |
cap = cv2.VideoCapture(video_path)
|
|
|
|
|
39 |
fourcc = cv2.VideoWriter_fourcc(*'mp4v')
|
40 |
output_path = "output_video.mp4"
|
41 |
+
out = cv2.VideoWriter(output_path, fourcc, 30.0, (int(cap.get(3)), int(cap.get(4))))
|
42 |
|
43 |
while cap.isOpened():
|
44 |
ret, frame = cap.read()
|
45 |
if not ret:
|
46 |
break
|
47 |
|
48 |
+
# Preprocess the frame
|
49 |
input_tensor = preprocess_frame(frame)
|
50 |
|
51 |
+
# Make predictions with the model
|
52 |
with torch.no_grad():
|
53 |
predictions = model(input_tensor)
|
54 |
|
55 |
+
# Convert the predictions back to an image format
|
56 |
output_frame = (predictions.squeeze().permute(1, 2, 0).numpy() * 255).astype(np.uint8)
|
|
|
57 |
|
58 |
+
# Write the processed frame to the output video
|
59 |
out.write(output_frame)
|
60 |
|
61 |
cap.release()
|
62 |
out.release()
|
63 |
return output_path
|
64 |
|
65 |
+
# Gradio interface function
|
66 |
def colorize_video(video):
|
67 |
model = load_model(MODEL_PATH)
|
68 |
+
output_video_path = process_video(model, video.name) # Use the video file name to read the video
|
69 |
return output_video_path
|
70 |
|
71 |
+
# Configure the Gradio interface
|
72 |
iface = gr.Interface(
|
73 |
fn=colorize_video,
|
74 |
+
inputs=gr.Video(label="Upload a black and white video"),
|
75 |
+
outputs=gr.Video(label="Colorized Video"),
|
76 |
+
title="Video Colorization",
|
77 |
+
description="Upload a black and white video to colorize it using the model."
|
78 |
)
|
79 |
|
80 |
if __name__ == '__main__':
|