Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -1,53 +1,152 @@
|
|
1 |
-
|
2 |
-
|
3 |
-
|
4 |
-
from torchvision import models
|
5 |
-
import gradio.inputs as gi
|
6 |
-
import gradio.outputs as go
|
7 |
import gradio as gr
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
8 |
|
9 |
-
#
|
10 |
-
class ResNet50(torch.nn.Module):
|
11 |
-
def __init__(self):
|
12 |
-
super(ResNet50, self).__init__()
|
13 |
-
self.resnet = models.resnet50(pretrained=True)
|
14 |
-
for param in self.resnet.parameters():
|
15 |
-
param.requires_grad = False
|
16 |
-
self.resnet.fc = torch.nn.Sequential(
|
17 |
-
torch.nn.Linear(2048, 2)
|
18 |
-
)
|
19 |
-
|
20 |
-
def forward(self, x):
|
21 |
-
x = self.resnet(x)
|
22 |
-
return x
|
23 |
-
|
24 |
-
# Load the pre-trained model
|
25 |
-
model = ResNet50()
|
26 |
-
model.load_state_dict(torch.load('best_modelv2.pth', map_location=torch.device('cpu')))
|
27 |
-
model.eval()
|
28 |
-
|
29 |
-
# Define transform for input images
|
30 |
-
data_transforms = transforms.Compose([
|
31 |
-
transforms.Resize((224, 224)),
|
32 |
-
transforms.ToTensor(),
|
33 |
-
transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
|
34 |
-
])
|
35 |
-
|
36 |
-
# Function to predict image label
|
37 |
-
def predict_image_label(image):
|
38 |
-
# Preprocess the image
|
39 |
-
image = data_transforms(image).unsqueeze(0)
|
40 |
-
|
41 |
-
# Make prediction
|
42 |
-
with torch.no_grad():
|
43 |
-
output = model(image)
|
44 |
-
_, predicted = torch.max(output, 1)
|
45 |
-
|
46 |
-
label = 'Leaf' if predicted.item() == 0 else 'Plant'
|
47 |
-
return label
|
48 |
-
|
49 |
-
# Create Gradio interface
|
50 |
-
# image = gi.Image(shape=(224, 224))
|
51 |
-
label = go.Label(num_top_classes=2)
|
52 |
-
|
53 |
-
gr.Interface(fn=predict_image_label,inputs="image", outputs=label, title="Leaf or Plant Classifier").launch()
|
|
|
1 |
+
|
2 |
+
|
3 |
+
|
|
|
|
|
|
|
4 |
import gradio as gr
|
5 |
+
from transformers import pipeline
|
6 |
+
|
7 |
+
# Load the model pipeline
|
8 |
+
pipe = pipeline("image-classification", "dima806/medicinal_plants_image_detection")
|
9 |
+
|
10 |
+
# Define the image classification function
|
11 |
+
def image_classifier(image):
|
12 |
+
# Perform image classification
|
13 |
+
outputs = pipe(image)
|
14 |
+
results = {}
|
15 |
+
for result in outputs:
|
16 |
+
results[result['label']] = result['score']
|
17 |
+
return results
|
18 |
+
|
19 |
+
# Define app title and description with HTML formatting
|
20 |
+
title = "<h1 style='text-align: center; color: #4CAF50;'>Image Classification</h1>"
|
21 |
+
description = "<p style='text-align: center; font-size: 18px;'>This application serves to classify skin lesion images based on their skin cancer type. Trained using Vision Transformer (ViT), it has achieved a validation accuracy of 86%.</p>"
|
22 |
+
|
23 |
+
# Define custom CSS styles for the Gradio app
|
24 |
+
custom_css = """
|
25 |
+
.gradio-interface {
|
26 |
+
max-width: 600px;
|
27 |
+
margin: auto;
|
28 |
+
border-radius: 10px;
|
29 |
+
box-shadow: 0px 0px 10px rgba(0, 0, 0, 0.1);
|
30 |
+
}
|
31 |
+
.title-container {
|
32 |
+
padding: 20px;
|
33 |
+
background-color: #f0f0f0;
|
34 |
+
border-top-left-radius: 10px;
|
35 |
+
border-top-right-radius: 10px;
|
36 |
+
}
|
37 |
+
.description-container {
|
38 |
+
padding: 20px;
|
39 |
+
}
|
40 |
+
"""
|
41 |
+
|
42 |
+
# Launch the Gradio interface with custom HTML and CSS
|
43 |
+
demo = gr.Interface(fn=image_classifier, inputs=gr.Image(type="pil"), outputs="label", title=title, description=description,
|
44 |
+
theme="gstaff/sketch", css=custom_css,
|
45 |
+
)
|
46 |
+
demo.launch()
|
47 |
+
|
48 |
+
|
49 |
+
|
50 |
+
|
51 |
+
|
52 |
+
|
53 |
+
|
54 |
+
|
55 |
+
|
56 |
+
|
57 |
+
|
58 |
+
|
59 |
+
|
60 |
+
|
61 |
+
|
62 |
+
|
63 |
+
|
64 |
+
|
65 |
+
|
66 |
+
|
67 |
+
|
68 |
+
|
69 |
+
|
70 |
+
|
71 |
+
|
72 |
+
|
73 |
+
|
74 |
+
|
75 |
+
|
76 |
+
|
77 |
+
|
78 |
+
|
79 |
+
|
80 |
+
|
81 |
+
|
82 |
+
|
83 |
+
|
84 |
+
|
85 |
+
|
86 |
+
|
87 |
+
|
88 |
+
|
89 |
+
|
90 |
+
|
91 |
+
|
92 |
+
|
93 |
+
|
94 |
+
|
95 |
+
|
96 |
+
|
97 |
+
|
98 |
+
|
99 |
+
|
100 |
+
# import torch
|
101 |
+
# from torchvision import transforms
|
102 |
+
# from PIL import Image
|
103 |
+
# from torchvision import models
|
104 |
+
# import gradio.inputs as gi
|
105 |
+
# import gradio.outputs as go
|
106 |
+
# import gradio as gr
|
107 |
+
|
108 |
+
# # Define the ResNet50 model
|
109 |
+
# class ResNet50(torch.nn.Module):
|
110 |
+
# def __init__(self):
|
111 |
+
# super(ResNet50, self).__init__()
|
112 |
+
# self.resnet = models.resnet50(pretrained=True)
|
113 |
+
# for param in self.resnet.parameters():
|
114 |
+
# param.requires_grad = False
|
115 |
+
# self.resnet.fc = torch.nn.Sequential(
|
116 |
+
# torch.nn.Linear(2048, 2)
|
117 |
+
# )
|
118 |
+
|
119 |
+
# def forward(self, x):
|
120 |
+
# x = self.resnet(x)
|
121 |
+
# return x
|
122 |
+
|
123 |
+
# # Load the pre-trained model
|
124 |
+
# model = ResNet50()
|
125 |
+
# model.load_state_dict(torch.load('best_modelv2.pth', map_location=torch.device('cpu')))
|
126 |
+
# model.eval()
|
127 |
+
|
128 |
+
# # Define transform for input images
|
129 |
+
# data_transforms = transforms.Compose([
|
130 |
+
# transforms.Resize((224, 224)),
|
131 |
+
# transforms.ToTensor(),
|
132 |
+
# transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
|
133 |
+
# ])
|
134 |
+
|
135 |
+
# # Function to predict image label
|
136 |
+
# def predict_image_label(image):
|
137 |
+
# # Preprocess the image
|
138 |
+
# image = data_transforms(image).unsqueeze(0)
|
139 |
+
|
140 |
+
# # Make prediction
|
141 |
+
# with torch.no_grad():
|
142 |
+
# output = model(image)
|
143 |
+
# _, predicted = torch.max(output, 1)
|
144 |
+
|
145 |
+
# label = 'Leaf' if predicted.item() == 0 else 'Plant'
|
146 |
+
# return label
|
147 |
+
|
148 |
+
# # Create Gradio interface
|
149 |
+
# # image = gi.Image(shape=(224, 224))
|
150 |
+
# label = go.Label(num_top_classes=2)
|
151 |
|
152 |
+
# gr.Interface(fn=predict_image_label,inputs="image", outputs=label, title="Leaf or Plant Classifier").launch()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|