VGG ARCHITECTURE EX -10

-
import tensorflow as tf
import matplotlib.pyplot as plt
from tensorflow.keras import datasets, layers, models, losses
(x_train, y_train), (x_test, y_test)=tf.keras.datasets.mnist.load_data()
x_train = tf.pad(x_train, [[0, 0], [2,2], [2,2]])/255
x_test = tf.pad(x_test, [[0, 0], [2,2], [2,2]])/255
x_train = tf.expand_dims(x_train, axis=3, name=None)
x_test = tf.expand_dims(x_test, axis=3, name=None)
x_val = x_train[-2000:,:,:,:]
y_val = y_train[-2000:]
x_train = x_train[:-2000,:,:,:]
y_train = y_train[:-2000]
model = models.Sequential([
 layers.experimental.preprocessing.Resizing(224, 224,
 interpolation="bilinear",
 input_shape=x_train.shape[1:]),
 layers.Conv2D(64, 3, strides=1, padding='same'),
 layers.Activation('relu'),
 layers.Conv2D(64, 3, strides=1, padding='same'),
 layers.Activation('relu'),
 layers.MaxPooling2D(2, strides=2),
 layers.Conv2D(128, 3, strides=1, padding='same'),
 layers.Activation('relu'),
 layers.Conv2D(128, 3, strides=1, padding='same'),
 layers.Activation('relu'),
 layers.MaxPooling2D(2, strides=2),
 layers.Conv2D(256, 3, strides=1, padding='same'),
 layers.Activation('relu'),
 layers.Conv2D(256, 1, strides=1, padding='same'),
 layers.Activation('relu'),
 layers.MaxPooling2D(2, strides=2),
 layers.Conv2D(512, 3, strides=1, padding='same'),
 layers.Activation('relu'),
 layers.Conv2D(512, 3, strides=1, padding='same'),
 layers.Activation('relu'),
 layers.Conv2D(512, 1, strides=1, padding='same'),
 layers.Activation('relu'),
 layers.MaxPooling2D(2, strides=2),
 layers.Conv2D(512, 3, strides=1, padding='same'),
 layers.Activation('relu'),
 layers.Conv2D(512, 3, strides=1, padding='same'),
 layers.Activation('relu'),
 layers.Conv2D(512, 1, strides=1, padding='same'),
 layers.Activation('relu'),
 layers.MaxPooling2D(2, strides=2),
 layers.Flatten(),
 layers.Dense(4096, activation='relu'),
 layers.Dropout(0.5),
 layers.Dense(4096, activation='relu'),
 layers.Dropout(0.5),
 layers.Dense(10, activation='softmax')
])
model.summary()
model.compile(optimizer='adam',
 loss=losses.sparse_categorical_crossentropy,
 metrics=['accuracy'])
history = model.fit(x_train, y_train,
 batch_size=64,
 epochs=20,
 validation_data=(x_val, y_val))
fig, axs = plt.subplots(2, 1, figsize=(15,15))
axs[0].plot(history.history['loss'])
axs[0].plot(history.history['val_loss'])
axs[0].title.set_text('Training Loss vs Validation Loss')
axs[0].legend(['Train', 'Val'])
axs[1].plot(history.history['accuracy'])
axs[1].plot(history.history['val_accuracy'])
axs[1].title.set_text('Training Accuracy vs Validation Accuracy')
axs[1].legend(['Train', 'Val'])
results = model.evaluate(x_test, y_test)
print("Loss = {}, Accuray = {}".format(results[0], results[1]))

Comments

Post a Comment

Popular posts from this blog

Naturally Simulated Prosthetic Arm With Electro Myo Grapy(EMG)-Robotic Arm (Product Creation):

-
Image
       In today’s society a hand amputee can expect a replacement hand that replicates a normal hand functions. However due to the cost constraints, the automated prosthetic arms are still out of the reach of needy people. So we made a naturally simulated prosthetic arm at cheap rate.  We used 3-D printer to print the parts of the arm and used servo motors to actuate the fingers, wrist and elbow. Electro Myo Graphy(EMG) sensor is used to get the signals from the muscle and controls the servo motors.  Graphic(EMG) sensor is used to get the signals from the muscle and controls the servo motor using Arduino. Structural View   Full assembly Servo assemble Servo reassembled for our application Designing Part Designing Part Main board and vibrating sensor for user alerts because when the user will touch Heat, Over weight Lifting the sensor will vibrate and give the sense. Labeling  Used connectors easy to maintenance Used connectors easy to maintenance Charging Port The Switch Is used for
Read more

IOT based Home automation Project Controlled by Google Assistant--Product Creation

-
Image
 Home Automation is controlling lighting, climate, entertainment systems, and appliances without a manual switch. It may also include home security such as access control and alarm systems. When connected and controlled with the Internet, home devices are an important constituent of the Internet of Things (“IoT”). NodeMCU is very popular in Home Automation. It’s WiFi capabilities and Arduino IDE support making it easier for IoT Applications. It is very tiny and has many Digital I/O pins, Serial Communication and I2C Communication. NodeMCU has a micro USB port to program it using your existing mobile cable (no additional programmer needed). There is a successor called ESP32 Development board which has more Analog pins and Digital pins.  Using Blynk is a mobile application which has its own server to process user requests. It is an open source application and anybody can use it in their Home Automation to control devices, monitor sensor data and get a notification by some trigger actions
Read more

SEQUENCE MODEL EX-13

-
  import tensorflow as tf from tensorflow import keras from tensorflow.keras import layers # Define Sequential model with 3 layers model = keras.Sequential([ layers.Dense( 2 , activation ="relu", name ="layer1"), layers.Dense( 3 , activation ="relu", name ="layer2"), layers.Dense( 4 , name ="layer3"), ]) # Call model on a test input x = tf .ones(( 3 , 3 )) y = model ( x ) # Create 3 layers layer1 = layers.Dense( 2 , activation ="relu", name ="layer1") layer2 = layers.Dense( 3 , activation ="relu", name ="layer2") layer3 = layers.Dense( 4 , name ="layer3") # Call layers on a test input x = tf .ones(( 3 , 3 )) y = layer3 ( layer2 ( layer1 ( x ))) model = keras.Sequential( [ layers.Dense( 2 , activation =&quot
Read more