Spaces:
Sleeping
Sleeping
Upload app.py
Browse files
app.py
ADDED
@@ -0,0 +1,99 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# Step 6.1: Define different input components
|
2 |
+
import gradio as gr
|
3 |
+
|
4 |
+
# g. define Dropdown data type
|
5 |
+
input_module1 = gr.inputs.Dropdown(choices=["KNN", "Softmax", "Deep Neural", "SVM", "Decision Tree","Random Forest","CNN"], label = "Model Selection")
|
6 |
+
# b. define image data type
|
7 |
+
input_module2 = gr.inputs.Dropdown(choices=["image1", "image2", "image3", "image4","image5"], label = "Sample Image")
|
8 |
+
|
9 |
+
# b. define image data type
|
10 |
+
output_module1 = gr.outputs.Textbox(label = "Predicted Class")
|
11 |
+
|
12 |
+
# b. define image data type
|
13 |
+
output_module2 = gr.outputs.Image(label = "Image")
|
14 |
+
|
15 |
+
# Step 6.3: Define a new function that accommodates the input modules.
|
16 |
+
def multi_inputs(input1, input2):
|
17 |
+
import numpy as np
|
18 |
+
import pickle
|
19 |
+
import matplotlib as mpl
|
20 |
+
import matplotlib.pyplot as plt
|
21 |
+
import tensorflow as tf
|
22 |
+
import scipy.ndimage.interpolation
|
23 |
+
(X_train_full, y_train_full), (X_test, y_test) = tf.keras.datasets.mnist.load_data()
|
24 |
+
|
25 |
+
X_valid, X_train = X_train_full[:5000], X_train_full[5000:]
|
26 |
+
y_valid, y_train = y_train_full[:5000], y_train_full[5000:]
|
27 |
+
|
28 |
+
if input2 == "image1":
|
29 |
+
img = X_test[1]
|
30 |
+
if input2 == "image2":
|
31 |
+
img = X_test[2]
|
32 |
+
if input2 == "image3":
|
33 |
+
img = X_test[3]
|
34 |
+
if input2 == "image4":
|
35 |
+
img = X_test[4]
|
36 |
+
if input2 == "image5":
|
37 |
+
img = X_test[5]
|
38 |
+
|
39 |
+
import matplotlib.pyplot as plt
|
40 |
+
import matplotlib.image as mpimg
|
41 |
+
#img = mpimg.imread(img)
|
42 |
+
#imag = scipy.ndimage.rotate(input2, 0, reshape=False, cval=0)
|
43 |
+
|
44 |
+
#class_names = ["T-shirt/top", "Trouser", "Pullover", "Dress", "Coat",
|
45 |
+
#"Sandal", "Shirt", "Sneaker", "Bag", "Ankle boot"]
|
46 |
+
test_shape = img.shape
|
47 |
+
test_width = test_shape[0]
|
48 |
+
test_height = test_shape[1]
|
49 |
+
|
50 |
+
image =img.reshape(test_width*test_height)
|
51 |
+
|
52 |
+
if input1 == "KNN":
|
53 |
+
loaded_model = pickle.load(open('KNN_Model.sav', 'rb'))
|
54 |
+
image = img.reshape((1, -1))
|
55 |
+
y_test_pred = loaded_model.predict(image)
|
56 |
+
|
57 |
+
|
58 |
+
if input1 == "Softmax":
|
59 |
+
loaded_model = pickle.load(open('Softmax_Model.sav', 'rb'))
|
60 |
+
image = img.reshape((1, -1))
|
61 |
+
y_test_pred = loaded_model.predict(image)
|
62 |
+
|
63 |
+
|
64 |
+
if input1 == "SVM":
|
65 |
+
loaded_model = pickle.load(open('svm_model.sav', 'rb'))
|
66 |
+
image = img.reshape((1, -1))
|
67 |
+
y_test_pred = loaded_model.predict(image)
|
68 |
+
|
69 |
+
if input1 == "Deep Neural":
|
70 |
+
loaded_model = pickle.load(open('dnn_model.sav', 'rb'))
|
71 |
+
image = img.reshape((1, -1))
|
72 |
+
y_test_pred = loaded_model.predict(image)
|
73 |
+
y_test_pred = np.argmax(y_test_pred,axis=1)
|
74 |
+
|
75 |
+
if input1 == "Decision Tree":
|
76 |
+
loaded_model = pickle.load(open('Decision_Tree_model.sav', 'rb'))
|
77 |
+
image = img.reshape((1, -1))
|
78 |
+
y_test_pred = loaded_model.predict(image)
|
79 |
+
|
80 |
+
if input1 == "Random Forest":
|
81 |
+
loaded_model = pickle.load(open('randomforest.sav', 'rb'))
|
82 |
+
image = img.reshape((1, -1))
|
83 |
+
y_test_pred = loaded_model.predict(image)
|
84 |
+
|
85 |
+
if input1 == "CNN":
|
86 |
+
loaded_model = pickle.load(open('CNN.sav', 'rb'))
|
87 |
+
img = img.reshape((1, -1))
|
88 |
+
y_test_pred = loaded_model.predict(img)
|
89 |
+
y_test_pred = np.argmax(y_test_pred,axis=1)
|
90 |
+
|
91 |
+
result = y_test_pred[0]
|
92 |
+
|
93 |
+
return result, img
|
94 |
+
|
95 |
+
# Step 6.4: Put all three component together into the gradio's interface function
|
96 |
+
gr.Interface(fn=multi_inputs,
|
97 |
+
inputs=[input_module1, input_module2],
|
98 |
+
outputs=[output_module1,output_module2]
|
99 |
+
).launch( debug = True)
|