Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -13,10 +13,12 @@ model = models.resnet50(weights=None)
|
|
13 |
# Revise fully connected layer to output 37 classes (num_classes = 37)
|
14 |
model.fc = torch.nn.Linear(2048, 37)
|
15 |
|
|
|
16 |
model.load_state_dict(torch.load('./resnet50_model_weights.pth', map_location=device))
|
17 |
|
18 |
model.eval()
|
19 |
|
|
|
20 |
transform = transforms.Compose([
|
21 |
transforms.Resize((224, 224)),
|
22 |
transforms.ToTensor(),
|
@@ -48,7 +50,7 @@ def classify_image(image):
|
|
48 |
probabilities, indices = torch.topk(probabilities, k=3)
|
49 |
# Return the class names with their corresponding probabilities
|
50 |
predictions = [(class_names[idx], prob.item()) for idx, prob in zip(indices[0], probabilities[0])]
|
51 |
-
return {class_name: prob for class_name, prob in predictions} # Return raw float numbers
|
52 |
|
53 |
# Path to the folder containing example images
|
54 |
examples_path = './examples'
|
@@ -66,22 +68,27 @@ examples = [[examples_path + "/" + img] for img in os.listdir(examples_path)]
|
|
66 |
# Create dropdown menu for users to see available classes (as reference, no direct connection to prediction)
|
67 |
dropdown = gr.Dropdown(choices=class_names, label="Recognizable Breeds", type="value")
|
68 |
|
69 |
-
#
|
70 |
-
demo = gr.Interface(
|
71 |
-
fn=classify_image,
|
72 |
-
inputs=[gr.Image(type="pil")], # Only image input is used for prediction
|
73 |
-
outputs=[gr.Label(num_top_classes=3, label="Top 3 Predictions")], # Outputs top 3 predictions with probabilities
|
74 |
-
examples=examples,
|
75 |
-
title='Oxford Pet ππ',
|
76 |
-
description='A ResNet50-based model for classifying 37 different pet breeds.',
|
77 |
-
article='[Oxford Project](https://github.com/Eric-Chung-0511/Learning-Record/tree/main/Data%20Science%20Projects/The%20Oxford-IIIT%20Pet%20Project)'
|
78 |
-
)
|
79 |
-
|
80 |
-
# Add dropdown to reference the recognizable breeds
|
81 |
with gr.Blocks() as demo_with_dropdown:
|
82 |
-
|
83 |
-
|
84 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
85 |
|
86 |
-
# Launch Gradio
|
87 |
demo_with_dropdown.launch()
|
|
|
13 |
# Revise fully connected layer to output 37 classes (num_classes = 37)
|
14 |
model.fc = torch.nn.Linear(2048, 37)
|
15 |
|
16 |
+
# Load Model weights
|
17 |
model.load_state_dict(torch.load('./resnet50_model_weights.pth', map_location=device))
|
18 |
|
19 |
model.eval()
|
20 |
|
21 |
+
# Transformation for the input image
|
22 |
transform = transforms.Compose([
|
23 |
transforms.Resize((224, 224)),
|
24 |
transforms.ToTensor(),
|
|
|
50 |
probabilities, indices = torch.topk(probabilities, k=3)
|
51 |
# Return the class names with their corresponding probabilities
|
52 |
predictions = [(class_names[idx], prob.item()) for idx, prob in zip(indices[0], probabilities[0])]
|
53 |
+
return {class_name: prob for class_name, prob in predictions} # Return raw float numbers # Return formatted percentages
|
54 |
|
55 |
# Path to the folder containing example images
|
56 |
examples_path = './examples'
|
|
|
68 |
# Create dropdown menu for users to see available classes (as reference, no direct connection to prediction)
|
69 |
dropdown = gr.Dropdown(choices=class_names, label="Recognizable Breeds", type="value")
|
70 |
|
71 |
+
# Use `gr.Blocks()` to define the full interface
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
72 |
with gr.Blocks() as demo_with_dropdown:
|
73 |
+
# Display markdown heading
|
74 |
+
gr.Markdown("# Oxford Pet ππ Recognizable Breeds")
|
75 |
+
|
76 |
+
# Dropdown as a reference for users
|
77 |
+
dropdown
|
78 |
+
|
79 |
+
# Image classification demo
|
80 |
+
gr.Image(label="Upload an image to classify").style(height="auto", width="100%")
|
81 |
+
|
82 |
+
# Gradio interface for the image input and label output
|
83 |
+
gr.Interface(
|
84 |
+
fn=classify_image,
|
85 |
+
inputs=gr.Image(type="pil"), # Only image input is used for prediction
|
86 |
+
outputs=gr.Label(num_top_classes=3, label="Top 3 Predictions"), # Outputs top 3 predictions with probabilities
|
87 |
+
examples=examples,
|
88 |
+
title='Oxford Pet ππ',
|
89 |
+
description='A ResNet50-based model for classifying 37 different pet breeds.',
|
90 |
+
article='[Oxford Project](https://github.com/Eric-Chung-0511/Learning-Record/tree/main/Data%20Science%20Projects/The%20Oxford-IIIT%20Pet%20Project)'
|
91 |
+
)
|
92 |
|
93 |
+
# Launch Gradio app
|
94 |
demo_with_dropdown.launch()
|