File size: 6,311 Bytes
affb669
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
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
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
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
import pandas as pd
import numpy as np
import tensorflow as tf
from keras.optimizers import SGD, Adagrad, RMSprop, Adadelta, Adam
from keras.models import Sequential
from keras.layers import Dense, Dropout
from keras.callbacks import EarlyStopping

import matplotlib.pyplot as plt
from matplotlib.backends.backend_agg import RendererAgg
_lock = RendererAgg.lock
from sklearn import linear_model, datasets
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score
from sklearn.tree import DecisionTreeClassifier

tf.random.set_seed(0)

import streamlit as st

data = datasets.load_breast_cancer()
df = pd.DataFrame(data["data"], columns=data["feature_names"])
df["target"] = data["target"]

X = df.drop("target", axis=1)
y = df["target"]
X_train, X_test, y_train, y_test = train_test_split(
    X, y, train_size=0.7, random_state=0
)


def plot_loss(history):
    fig, ax = plt.subplots()
    ax.plot(history.history["loss"], label="loss")
    ax.plot(history.history["val_loss"], label="val_loss")
    ax.set_xlabel("Epoch")
    ax.set_ylabel("Error")
    ax.set_title("Train Loss vs Validation Loss")
    ax.legend()
    ax.grid(True)
    return fig


def create_model():
    model = Sequential()
    model.add(
        Dense(32, kernel_initializer="normal", input_dim=30, activation="leaky_relu")
    )
    model.add(Dense(16, kernel_initializer="uniform", activation="leaky_relu"))
    model.add(Dropout(rate=0.3))
    model.add(Dense(16, kernel_initializer="uniform", activation="sigmoid"))
    model.add(Dropout(rate=0.4))
    model.add(Dense(1, activation="sigmoid"))
    return model


def fit_model(model, optmizer, X_train, X_test, y_test, batch_size=32):
    model.compile(loss="binary_crossentropy", optimizer=optimizer, metrics=["accuracy"])
    callback = EarlyStopping(monitor="loss", patience=10)
    history = model.fit(
        X_train,
        y_train,
        # Setting Batch Size to number of samples for Vanilla GD
        batch_size=X_train.shape[0],
        validation_data=(X_test, y_test),
        epochs=150,
        callbacks=[callback],
        verbose=0,
    )
    return history


gd_types = [
    "Gradient Descent",
    "Stochastic Gradient Descent",
    "Mini-Batch Gradient Descent",
    "Adagrad",
    "RMSProp",
    "Adam",
]

with st.sidebar:
    choice = st.selectbox("Optimizer:", options=gd_types)

    if (
        choice == "Gradient Descent"
        or choice == "Stochastic Gradient Descent"
        or choice == "Mini-Batch Gradient Descent"
    ):
        lr = st.slider(
            "Learning Rate:", min_value=0.01, max_value=1.00, value=0.01, step=0.01
        )
        if choice == "Mini-Batch Gradient Descent":
            batch_size = st.slider(
                "Batch Size:", min_value=1, max_value=100, value=50, step=10
            )
        else:
            batch_size = st.select_slider(
                "Batch Size:", [1, 2, 4, 8, 16, 32, 64], disabled=True
            )
        momentum = st.slider(
            "Momentum Factor:", min_value=0.01, max_value=1.00, value=0.01, step=0.01
        )
        nag = st.checkbox("Nesterov Accelerated Momentum")

    elif choice == "Adagrad":
        lr = st.slider(
            "Learning Rate:", min_value=0.01, max_value=1.00, value=0.1, step=0.01
        )
        batch_size = st.slider(
            "Batch Size:", min_value=1, max_value=100, value=50, step=10
        )

    elif choice == "RMSProp":
        lr = st.slider(
            "Learning Rate:", min_value=0.01, max_value=1.00, value=0.01, step=0.01
        )
        batch_size = st.slider(
            "Batch Size:", min_value=1, max_value=100, value=50, step=10
        )
        rho = st.slider(
            "Exponential Decay Rate:", min_value=0.1, max_value=1.0, value=0.9, step=0.1
        )

    elif choice == "Adam":
        lr = st.slider(
            "Learning Rate:", min_value=0.01, max_value=1.00, value=0.01, step=0.01
        )
        batch_size = st.slider(
            "Batch Size:", min_value=1, max_value=100, value=50, step=10
        )
        beta1 = st.slider(
            "Exponential Decay Rate for Moments:",
            min_value=0.1,
            max_value=1.0,
            value=0.9,
            step=0.1,
        )
        beta2 = st.slider(
            "Exponential Decay Rate for Variance:",
            min_value=0.01,
            max_value=1.00,
            value=0.99,
            step=0.01,
        )

st.title("Optimizers in Deep Learning")

st.write(
    "A Neural Network has been trained on the Breast Cancer Dataset. We monitor the convergence of different optimizers during the training."
)

st.subheader(choice)

if choice == "Gradient Descent":
    model = create_model()
    optimizer = SGD(learning_rate=lr, momentum=momentum, nesterov=nag)
    history = fit_model(
        model, optimizer, X_train, X_test, y_test, batch_size=X_train.shape[0]
    )
    st.pyplot(plot_loss(history))

elif choice == "Stochastic Gradient Descent":
    model = create_model()
    optimizer = SGD(learning_rate=lr, momentum=momentum, nesterov=nag)
    history = fit_model(model, optimizer, X_train, X_test, y_test, batch_size=1)
    st.pyplot(plot_loss(history))

elif choice == "Mini-Batch Gradient Descent":
    model = create_model()
    optimizer = SGD(learning_rate=lr, momentum=momentum, nesterov=nag)
    history = fit_model(
        model, optimizer, X_train, X_test, y_test, batch_size=batch_size
    )
    st.pyplot(plot_loss(history))

elif choice == "Adagrad":
    model = create_model()
    optimizer = Adagrad(learning_rate=lr)
    history = fit_model(
        model, optimizer, X_train, X_test, y_test, batch_size=batch_size
    )
    st.pyplot(plot_loss(history))

elif choice == "RMSProp":
    model = create_model()
    optimizer = RMSprop(learning_rate=lr, rho=rho)
    history = fit_model(
        model, optimizer, X_train, X_test, y_test, batch_size=batch_size
    )
    st.pyplot(plot_loss(history))

elif choice == "Adam":
    model = create_model()
    optimizer = Adam(learning_rate=lr, beta_1=beta1, beta_2=beta2)
    history = fit_model(
        model, optimizer, X_train, X_test, y_test, batch_size=batch_size
    )
    st.pyplot(plot_loss(history))


st.write("The dataset can be viewed below:")

st.dataframe(data=df, width=1000, height=200)