javiferfer

import gzip
import numpy as np
import matplotlib.pyplot as plt
import struct
import time
import torch
import torchvision
from torch import nn
import torch.nn.functional as F
from torch.utils.tensorboard import SummaryWriter

javiferfer

accuracy = history.history['accuracy']
val_accuracy = history.history['val_accuracy']
loss = history.history['loss']
val_loss = history.history['val_loss']

javiferfer

def evaluate_score(model, test_x, test_y):
    score = model.evaluate(test_x,test_y,verbose=0)
    print('Test Loss : {:.4f}'.format(score[0]))
    print('Test Accuracy : {:.4f}'.format(score[1]))
    
evaluate_score(model, x_test, y_test)

javiferfer

history = compile_model(model, x_train, y_train, x_validate, y_validate, batch_size=512, no_epochs=15)

javiferfer

def load_trained_model(filepath):
    new_model = load_model(filepath)
    return model
    
model = load_trained_model("models/best_model.hdf5")

javiferfer

def run_model(model, train_x, train_y, validate_x, validate_y, batch_size, no_epochs):
    
    datetime_now = datetime.datetime.now()
    datetime_now = datetime_now.strftime("%Y_%m_%d_%H_%M_%S")
    print(datetime_now)
    os.mkdir('models/' + str(datetime_now))
    
    filepath = 'models/' + str(datetime_now) + '/model-{epoch:02d}-{val_accuracy:.2f}.hdf5'
    checkpoint = ModelCheckpoint(filepath, monitor='loss', verbose=1,
                                 save_best_only=False, mode='auto', period=1)

javiferfer

def keras_cnn(input_shape, num_classes):
    model = Sequential()
    model.add(Conv2D(filters=32, kernel_size=3, activation='relu', input_shape = input_shape))
    model.add(MaxPooling2D(pool_size=2) )
    model.add(Conv2D(filters=64, kernel_size=3, activation='relu'))
    model.add(Dropout(0.2))
    model.add(Flatten())
    model.add(Dense(32, activation='relu'))
    model.add(Dense(num_classes, activation = 'softmax'))
   

javiferfer

y_train = keras.utils.to_categorical(train_y, num_classes)
y_test = keras.utils.to_categorical(test_y, num_classes)

x_train, x_validate, y_train, y_validate = train_test_split(train_x, y_train, test_size = 0.2, random_state = 16)

input_shape = (28, 28, 1)
x_train = x_train.reshape(x_train.shape[0], *input_shape)
x_validate = x_validate.reshape(x_validate.shape[0], *input_shape)
x_test = test_x.reshape(test_x.shape[0], *input_shape)

javiferfer

num_classes = 10
target_names = ['top', 'trouser', 'pullover', 'dress', 'coat', 'sandal', 'shirt', 'sneaker', 'bag', 'boot']

javiferfer

train_x = train_x/255
test_x = test_x/255