Implementation of ICNet
In this notebook, an implementation of ICNet is presented which is an architecture which uses a trade-off between complexity and inference time efficiently. The architecture is evaluated against the Oxford pets dataset. This notebook has reused material from the Image Segmentation Tutorial on Tensorflow
Importing the required packages.
import tensorflow as tf
import tensorflow_datasets as tfds
import matplotlib.pyplot as plt
from tensorflow.keras.layers import *
from tensorflow.keras.models import *
import numpy as np
from tensorflow.keras.applications.resnet50 import ResNet50
import matplotlib.pyplot as plt
import matplotlib.image as mpimg
import tensorflow_addons as tfa
from hyperopt import fmin, tpe, hp, Trials, STATUS_OK, SparkTrials
Loading and transforming the dataset.
def normalize(input_image, input_mask):
input_image = tf.cast(input_image, tf.float32) / 255.0
input_mask -= 1
return input_image, input_mask
# Function for resizing the train images to the desired input shape of HxW as well as augmenting the training images.
def load_image_train_noTf(datapoint, wanted_height: int, wanted_width: int):
input_image = tf.image.resize(datapoint['image'], (wanted_height, wanted_width))
input_mask = tf.image.resize(datapoint['segmentation_mask'], (wanted_height, wanted_width))
if tf.random.uniform(()) > 0.5:
input_image = tf.image.flip_left_right(input_image)
input_mask = tf.image.flip_left_right(input_mask)
input_image, input_mask = normalize(input_image, input_mask)
input_mask = tf.math.round(input_mask)
return input_image, input_mask
# Function for resizing the test images to the desired output shape (no augmenation).
def load_image_test(datapoint, wanted_height: int, wanted_width: int):
input_image = tf.image.resize(datapoint['image'], (wanted_height, wanted_width))
input_mask = tf.image.resize(datapoint['segmentation_mask'], (wanted_height, wanted_width))
input_image, input_mask = normalize(input_image, input_mask)
return input_image, input_mask
# Functions for resizing the image to the desired size of factor 2 or 4 to be inputted to the ICNet architecture.
def resize_image16(img, mask, wanted_height: int, wanted_width: int):
input_image = tf.image.resize(img, (wanted_height//16, wanted_width//16))
input_mask=tf.image.resize(mask, (wanted_height//16, wanted_width//16))
input_mask = tf.math.round(input_mask)
return input_image, input_mask
def resize_image8(img, mask, wanted_height: int, wanted_width: int):
input_image = tf.image.resize(img, (wanted_height//8, wanted_width//8))
input_mask=tf.image.resize(mask, (wanted_height//8, wanted_width//8))
input_mask = tf.math.round(input_mask)
return input_image, input_mask
def resize_image4(img, mask, wanted_height: int, wanted_width: int):
input_image = tf.image.resize(img, (wanted_height//4, wanted_width//4))
input_mask=tf.image.resize(mask, (wanted_height//4, wanted_width//4))
input_mask = tf.math.round(input_mask)
return input_image, input_mask
def create_datasets(wanted_height:int, wanted_width:int, BATCH_SIZE:int = 64, BUFFER_SIZE:int = 1000):
#Loading the dataset
dataset, info = tfds.load('oxford_iiit_pet:3.*.*', with_info=True)
n_train = info.splits['train'].num_examples
n_test = info.splits['test'].num_examples
#Creating the ndarray in the correct shapes for training data
train_original_img = np.ndarray(shape=(n_train, wanted_height, wanted_width, 3), dtype=np.float32)
train_original_mask = np.ndarray(shape=(n_train, wanted_height, wanted_width, 1), dtype=np.float32)
train16_mask = np.ndarray(shape=(n_train, wanted_height//16, wanted_width//16, 1), dtype=np.float32)
train8_mask = np.ndarray(shape=(n_train, wanted_height//8, wanted_width//8, 1), dtype=np.float32)
train4_mask = np.ndarray(shape=(n_train, wanted_height//4, wanted_width//4, 1), dtype=np.float32)
#Loading the data into the arrays
count = 0
for datapoint in dataset['train']:
img_orig, mask_orig = load_image_train_noTf(datapoint, wanted_height, wanted_width)
train_original_img[count]=img_orig
train_original_mask[count]=mask_orig
img16, mask16 = resize_image16(img_orig, mask_orig, wanted_height, wanted_width)
train16_mask[count]=(mask16)
img8, mask8 = resize_image8(img_orig, mask_orig, wanted_height, wanted_width)
train8_mask[count]=(mask8)
img4, mask4 = resize_image4(img_orig, mask_orig, wanted_height, wanted_width)
train4_mask[count]=(mask4)
count+=1
#Creating the ndarrays in the correct shapes for test data
test_original_img = np.ndarray(shape=(n_test,wanted_height,wanted_width,3), dtype=np.float32)
test_original_mask = np.ndarray(shape=(n_test,wanted_height,wanted_width,1), dtype=np.float32)
test16_mask = np.ndarray(shape=(n_test,wanted_height//16,wanted_width//16,1), dtype=np.float32)
test8_mask = np.ndarray(shape=(n_test,wanted_height//8,wanted_width//8,1), dtype=np.float32)
test4_mask = np.ndarray(shape=(n_test,wanted_height//4,wanted_width//4,1), dtype=np.float32)
#Loading the data into the arrays
count=0
for datapoint in dataset['test']:
img_orig, mask_orig = load_image_test(datapoint, wanted_height, wanted_width)
test_original_img[count]=(img_orig)
test_original_mask[count]=(mask_orig)
img16, mask16 = resize_image16(img_orig, mask_orig, wanted_height, wanted_width)
test16_mask[count]=(mask16)
img8, mask8 = resize_image8(img_orig, mask_orig, wanted_height, wanted_width)
test8_mask[count]=(mask8)
img4, mask4 = resize_image4(img_orig, mask_orig, wanted_height, wanted_width)
test4_mask[count]=(mask4)
count+=1
train_dataset = [train_original_img, train16_mask, train8_mask, train4_mask, train_original_mask]
orig_test_dataset = [test_original_img, test16_mask, test8_mask, test4_mask, test_original_mask]
return train_dataset, orig_test_dataset, train_original_mask[0], train_original_img[0], n_train, n_test
Running and loading the functions to create and save the transformed data.
train_dataset, orig_test_dataset, sample_mask, sample_image, n_train, n_test = create_datasets(128,128, 64, 1000)
Defining the function for displaying images and the model's predictions jointly.
def display(display_list):
plt.figure(figsize=(15, 15))
title = ['Input Image', 'True Mask', 'Predicted Mask']
for i in range(len(display_list)):
plt.subplot(1, len(display_list), i+1)
plt.title(title[i])
plt.imshow(tf.keras.preprocessing.image.array_to_img(display_list[i]))
plt.axis('off')
plt.show()
sample_image, sample_mask = sample_image, sample_mask
display([sample_image, sample_mask])
Defining the functions needed for the PSPNet module.
# Function for the pooling module which takes the output of ResNet50 as input as well as its width and height and pool it with a factor.
def pool_block(cur_tensor,
image_width,
image_height,
pooling_factor,
activation):
strides = [int(np.round(float(image_width)/pooling_factor)),
int(np.round(float(image_height)/pooling_factor))]
pooling_size = strides
x = AveragePooling2D(pooling_size, strides=strides, padding='same')(cur_tensor)
x = Conv2D(128,(1,1),padding='same')(x)
x = BatchNormalization()(x)
x = Activation(activation)(x)
x = tf.keras.layers.experimental.preprocessing.Resizing(
image_height, image_width, interpolation="bilinear")(x) # Resizing images to correct shape for future concat
return x
# Function for formatting the resnet model to a modified one which takes advantage of dilation rates instead of strides in the final blocks.
def modify_ResNet_Dilation(model):
for i in range(0,4):
model.get_layer('conv4_block1_{}_conv'.format(i)).strides = 1
model.get_layer('conv4_block1_{}_conv'.format(i)).dilation_rate = 2
model.get_layer('conv5_block1_{}_conv'.format(i)).strides = 1
model.get_layer('conv5_block1_{}_conv'.format(i)).dilation_rate = 4
model.save('/tmp/my_model')
new_model = tf.keras.models.load_model('/tmp/my_model')
return new_model
# Function for creating the PSPNet model. The inputs is the number of classes to classify, number of filters to use, kernel_size, activation function,
# input image width and height and a boolean for knowing if the module is part of the ICNet or not.
def PSPNet(n_classes: int,
n_filters: int,
kernel_size: tuple,
activation: str,
image_width: int,
image_height: int,
isICNet: bool = False,
dropout: bool = True,
bn: bool = True
):
if isICNet:
input_shape=(None, None, 3)
else:
input_shape=(image_height,image_width,3)
encoder=ResNet50(include_top=False, weights='imagenet', input_shape=input_shape)
encoder=modify_ResNet_Dilation(encoder)
#encoder.trainable=False
resnet_output=encoder.output
pooling_layer=[]
pooling_layer.append(resnet_output)
output=(resnet_output)
h = image_height//8
w = image_width//8
for i in [1,2,3,6]:
pool = pool_block(output, h, w, i, activation)
pooling_layer.append(pool)
concat=Concatenate()(pooling_layer)
output_layer=Conv2D(filters=n_classes, kernel_size=(1,1), padding='same')(concat)
final_layer=UpSampling2D(size=(8,8), data_format='channels_last', interpolation='bilinear')(output_layer)
final_model=tf.keras.models.Model(inputs=encoder.input, outputs=final_layer)
return final_model
#model = PSPNet(3, 16, (3,3), 'relu', 128,128)
Defining the functions needed for the ICNet.
# Function for adding stage 4 and 5 of ResNet50 to the 1/4 image size branch of the ICNet.
def PSP_rest(input_prev: tf.Tensor):
y_ = input_prev
#Stage 4
#Conv_Block
y = Conv2D(256, 1, dilation_rate=2, padding='same', name='C4_block1_conv1')(y_)
y = BatchNormalization(name='C4_block1_bn1')(y)
y = Activation('relu', name='C4_block1_act1')(y)
y = Conv2D(256, 3, dilation_rate=2, padding='same', name='C4_block1_conv2')(y)
y = BatchNormalization(name='C4_block1_bn2')(y)
y = Activation('relu', name='C4_block1_act2')(y)
y_ = Conv2D(1024, 1, dilation_rate=2, padding='same', name='C4_block1_conv0')(y_)
y = Conv2D(1024, 1, dilation_rate=2, padding='same', name='C4_block1_conv3')(y)
y_ = BatchNormalization(name='C4_block1_bn0')(y_)
y = BatchNormalization(name='C4_block1_bn3')(y)
y = Add(name='C4_skip1')([y_,y])
y_ = Activation('relu', name='C4_block1_act3')(y)
#IDBLOCK1
y = Conv2D(256, 1, dilation_rate=2, padding='same', name='C4_block2_conv1')(y_)
y = BatchNormalization(name='C4_block2_bn1')(y)
y = Activation('relu', name='C4_block2_act1')(y)
y = Conv2D(256, 3, dilation_rate=2, padding='same', name='C4_block2_conv2')(y)
y = BatchNormalization(name='C4_block2_bn2')(y)
y = Activation('relu', name='C4_block2_act2')(y)
y = Conv2D(1024,1, dilation_rate=2, padding='same', name='C4_block2_conv3')(y)
y = BatchNormalization(name='C4_block2_bn3')(y)
y = Add(name='C4_skip2')([y_,y])
y_ = Activation('relu', name='C4_block2_act3')(y)
#IDBLOCK2
y = Conv2D(256, 1, dilation_rate=2, padding='same', name='C4_block3_conv1')(y_)
y = BatchNormalization(name='C4_block3_bn1')(y)
y = Activation('relu', name='C4_block3_act1')(y)
y = Conv2D(256, 3, dilation_rate=2, padding='same', name='C4_block3_conv2')(y)
y = BatchNormalization(name='C4_block3_bn2')(y)
y = Activation('relu', name='C4_block3_act2')(y)
y = Conv2D(1024,1, dilation_rate=2, padding='same', name='C4_block3_conv3')(y)
y = BatchNormalization(name='C4_block3_bn3')(y)
y = Add(name='C4_skip3')([y_,y])
y_ = Activation('relu', name='C4_block3_act3')(y)
#IDBlock3
y = Conv2D(256, 1, dilation_rate=2, padding='same', name='C4_block4_conv1')(y_)
y = BatchNormalization(name='C4_block4_bn1')(y)
y = Activation('relu', name='C4_block4_act1')(y)
y = Conv2D(256, 3, dilation_rate=2, padding='same', name='C4_block4_conv2')(y)
y = BatchNormalization(name='C4_block4_bn2')(y)
y = Activation('relu', name='C4_block4_act2')(y)
y = Conv2D(1024,1, dilation_rate=2, padding='same', name='C4_block4_conv3')(y)
y = BatchNormalization(name='C4_block4_bn3')(y)
y = Add(name='C4_skip4')([y_,y])
y_ = Activation('relu', name='C4_block4_act3')(y)
#ID4
y = Conv2D(256, 1, dilation_rate=2, padding='same', name='C4_block5_conv1')(y_)
y = BatchNormalization(name='C4_block5_bn1')(y)
y = Activation('relu', name='C4_block5_act1')(y)
y = Conv2D(256, 3, dilation_rate=2, padding='same', name='C4_block5_conv2')(y)
y = BatchNormalization(name='C4_block5_bn2')(y)
y = Activation('relu', name='C4_block5_act2')(y)
y = Conv2D(1024,1, dilation_rate=2, padding='same', name='C4_block5_conv3')(y)
y = BatchNormalization(name='C4_block5_bn3')(y)
y = Add(name='C4_skip5')([y_,y])
y_ = Activation('relu', name='C4_block5_act3')(y)
#ID5
y = Conv2D(256, 1, dilation_rate=2, padding='same', name='C4_block6_conv1')(y_)
y = BatchNormalization(name='C4_block6_bn1')(y)
y = Activation('relu', name='C4_block6_act1')(y)
y = Conv2D(256, 3, dilation_rate=2, padding='same', name='C4_block6_conv2')(y)
y = BatchNormalization(name='C4_block6_bn2')(y)
y = Activation('relu', name='C4_block6_act2')(y)
y = Conv2D(1024,1, dilation_rate=2, padding='same', name='C4_block6_conv3')(y)
y = BatchNormalization(name='C4_block6_bn3')(y)
y = Add(name='C4_skip6')([y_,y])
y_ = Activation('relu', name='C4_block6_act3')(y)
#Stage 5
#Conv
y = Conv2D(512, 1, dilation_rate=4,padding='same', name='C5_block1_conv1')(y_)
y = BatchNormalization(name='C5_block1_bn1')(y)
y = Activation('relu', name='C5_block1_act1')(y)
y = Conv2D(512, 3, dilation_rate=4,padding='same', name='C5_block1_conv2')(y)
y = BatchNormalization(name='C5_block1_bn2')(y)
y = Activation('relu', name='C5_block1_act2')(y)
y_ = Conv2D(2048, 1, dilation_rate=4,padding='same', name='C5_block1_conv0')(y_)
y = Conv2D(2048, 1, dilation_rate=4,padding='same', name='C5_block1_conv3')(y)
y_ = BatchNormalization(name='C5_block1_bn0')(y_)
y = BatchNormalization(name='C5_block1_bn3')(y)
y = Add(name='C5_skip1')([y_,y])
y_ = Activation('relu', name='C5_block1_act3')(y)
#ID
y = Conv2D(512, 1, dilation_rate=4,padding='same', name='C5_block2_conv1')(y_)
y = BatchNormalization(name='C5_block2_bn1')(y)
y = Activation('relu', name='C5_block2_act1')(y)
y = Conv2D(512, 3, dilation_rate=4,padding='same', name='C5_block2_conv2')(y)
y = BatchNormalization(name='C5_block2_bn2')(y)
y = Activation('relu', name='C5_block2_act2')(y)
y = Conv2D(2048, 1, dilation_rate=4,padding='same', name='C5_block2_conv3')(y)
y = BatchNormalization(name='C5_block2_bn3')(y)
y = Add(name='C5_skip2')([y_,y])
y_ = Activation('relu', name='C5_block2_act3')(y)
#ID
y = Conv2D(512, 1, dilation_rate=4,padding='same', name='C5_block3_conv1')(y_)
y = BatchNormalization(name='C5_block3_bn1')(y)
y = Activation('relu', name='C5_block3_act1')(y)
y = Conv2D(512, 3, dilation_rate=4,padding='same', name='C5_block3_conv2')(y)
y = BatchNormalization(name='C5_block3_bn2')(y)
y = Activation('relu', name='C5_block3_act2')(y)
y = Conv2D(2048, 1, dilation_rate=4,padding='same', name='C5_block3_conv3')(y)
y = BatchNormalization(name='C5_block3_bn3')(y)
y = Add(name='C5_skip3')([y_,y])
y_ = Activation('relu', name='C5_block3_act3')(y)
return(y_)
# Function for the CFF module in the ICNet architecture. The inputs are which stage (1 or 2), the output from the smaller branch, the output from the
# larger branch, n_classes and the width and height of the output of the smaller branch.
def CFF(stage: int, F_small, F_large, n_classes: int, input_width_small: int, input_height_small: int):
F_up = tf.keras.layers.experimental.preprocessing.Resizing(int(input_width_small*2), int(input_height_small*2), interpolation="bilinear", name="Upsample_x2_small_{}".format(stage))(F_small)
F_aux = Conv2D(n_classes, 1, name="CC_{}".format(stage), activation='softmax')(F_up)
intermediate_f_small = Conv2D(128, 3, dilation_rate=2, padding='same', name="intermediate_f_small_{}".format(stage))(F_up)
intermediate_f_small_bn = BatchNormalization(name="intermediate_f_small_bn_{}".format(stage))(intermediate_f_small)
intermediate_f_large = Conv2D(128, 1, padding='same', name="intermediate_f_large_{}".format(stage))(F_large)
intermediate_f_large_bn = BatchNormalization(name="intermediate_f_large_bn_{}".format(stage))(intermediate_f_large)
intermediate_f_sum = Add(name="add_intermediates_{}".format(stage))([intermediate_f_small_bn,intermediate_f_large_bn])
intermediate_f_relu = Activation('relu', name="activation_CFF_{}".format(stage))(intermediate_f_sum)
return F_aux, intermediate_f_relu
# Function for the high-res branch of ICNet where image is in scale 1:1. The inputs are the input image, number of filters, kernel size and desired activation function.
def ICNet_1(input_obj,
n_filters: int,
kernel_size: tuple,
activation: str):
temp=input_obj
for i in range(1,4):
conv1=Conv2D(filters=n_filters*2*i, kernel_size=kernel_size, strides=(2,2), padding='same')(temp)
batch_norm1=BatchNormalization()(conv1)
temp=Activation(activation)(batch_norm1)
return temp
# Function for creating the ICNet model. The inputs are the width and height of the images to be used by the model, number of classes, number of filters, kernel size and
# desired activation function.
def ICNet(image_width: int,
image_height: int,
n_classes: int,
n_filters: int = 16,
kernel_size: tuple = (3,3),
activation: str = 'relu'):
input_shape=[image_width,image_height,3]
input_obj = tf.keras.Input(shape=input_shape, name="input_img_1")
input_obj_4 = tf.keras.layers.experimental.preprocessing.Resizing(
image_width//4, image_height//4, interpolation="bilinear", name="input_img_4")(input_obj)
input_obj_2 = tf.keras.layers.experimental.preprocessing.Resizing(
image_width//2, image_height//2, interpolation="bilinear", name="input_img_2")(input_obj)
ICNet_Model1=ICNet_1(input_obj, n_filters, kernel_size, activation)
PSPModel = PSPNet(n_classes, n_filters, kernel_size, activation, image_width//4, image_height//4, True)
last_layer = PSPModel.get_layer('conv4_block3_out').output
PSPModel_2_4 = tf.keras.models.Model(inputs=PSPModel.input, outputs=last_layer, name="JointResNet_2_4")
ICNet_Model4 = PSPModel_2_4(input_obj_4)
ICNet_Model2 = PSPModel_2_4(input_obj_2)
ICNet_4_rest = PSP_rest(ICNet_Model4)
out1, last_layer = CFF(1, ICNet_4_rest, ICNet_Model2, n_classes, image_width//32, image_height//32)
out2, last_layer = CFF(2, last_layer, ICNet_Model1, n_classes, image_width//16, image_height//16)
upsample_2 = UpSampling2D(2, interpolation='bilinear', name="Upsampling_final_prediction")(last_layer)
output = Conv2D(n_classes, 1, name="CC_fin", activation='softmax')(upsample_2)
final_output = UpSampling2D(4, interpolation='bilinear', name='final_output')(output)
final_model = tf.keras.models.Model(inputs=input_obj, outputs=[out1, out2, output, final_output])
return final_model
We create a callback for early stopping to prevent overfitting.
early_stopping = tf.keras.callbacks.EarlyStopping(
monitor='val_final_output_loss', patience=4, verbose=0
)
Finally, we fit the model to the Oxford dataset.
def batch_generator(batch_size):
indices = np.arange(len(train_dataset[0]))
batch=[]
while True:
# it might be a good idea to shuffle your data before each epoch
np.random.shuffle(indices)
for i in indices:
batch.append(i)
if len(batch)==batch_size:
yield train_dataset[0][batch], {'CC_1': train_dataset[1][batch], 'CC_2': train_dataset[2][batch], 'CC_fin': train_dataset[3][batch], 'final_output': train_dataset[4][batch]}
batch=[]
def batch_generator_eval(batch_size):
indices = np.arange(len(train_dataset[0]))
batch=[]
while True:
for i in indices:
batch.append(i)
if len(batch)==batch_size:
yield orig_test_dataset[0][batch], {'CC_1': orig_test_dataset[1][batch], 'CC_2': orig_test_dataset[2][batch], 'CC_fin': orig_test_dataset[3][batch], 'final_output': orig_test_dataset[4][batch]}
batch=[]
def train(params):
VAL_SUBSPLITS=5
EPOCHS=20
VALIDATION_STEPS = n_test//params['batch_size']//VAL_SUBSPLITS
STEPS_PER_EPOCH = n_train // params['batch_size']
BATCH_SIZE = params['batch_size']
"""
An example train method that calls into HorovodRunner.
This method is passed to hyperopt.fmin().
:param params: hyperparameters. Its structure is consistent with how search space is defined. See below.
:return: dict with fields 'loss' (scalar loss) and 'status' (success/failure status of run)
"""
model=ICNet(128,128,3)
model.compile(optimizer=tfa.optimizers.AdamW(learning_rate=params['learning_rate'], weight_decay=0.0001),
loss=tf.keras.losses.SparseCategoricalCrossentropy(), loss_weights=[0.4,0.4,1,0],
metrics="acc")
train_dataset_temp = batch_generator(BATCH_SIZE)
test_dataset = batch_generator_eval(BATCH_SIZE)
model_history = model.fit(train_dataset_temp, epochs=EPOCHS, steps_per_epoch=STEPS_PER_EPOCH, callbacks=[early_stopping])
loss = model.evaluate(test_dataset, steps=VALIDATION_STEPS)[4]
del model, train_dataset_temp, test_dataset, model_history
tf.keras.backend.clear_session()
return {'loss': loss, 'status': STATUS_OK}
import numpy as np
space = {
'learning_rate': hp.loguniform('learning_rate', np.log(1e-4), np.log(1e-1)),
'batch_size': hp.choice('batch_size', [32, 64, 128]),
}
import mlflow.tensorflow
algo=tpe.suggest
mlflow.tensorflow.autolog(every_n_iter=1)
spark_trials = SparkTrials(parallelism=2)
best_param = fmin(
fn=train,
space=space,
algo=algo,
max_evals=8,
return_argmin=False,
trials = spark_trials,
)
print(best_param)
Without Spark Trials: 1 hour
With Spark Trials: 33.18 min
We visualize the accuracies and losses through the library matplotlib.
loss = model_history.history['loss']
acc = model_history.history['final_output_acc']
val_loss = model_history.history['val_loss']
val_loss1 = model_history.history['val_CC_1_loss']
val_loss2 = model_history.history['val_CC_2_loss']
val_loss3 = model_history.history['val_CC_fin_loss']
val_loss4 = model_history.history['val_final_output_loss']
val_acc1 = model_history.history['val_CC_1_acc']
val_acc2 = model_history.history['val_CC_2_acc']
val_acc3 = model_history.history['val_CC_fin_acc']
val_acc4 = model_history.history['val_final_output_acc']
epochs = range(16)
plt.figure(figsize=(20,3))
plt.subplot(1,4,1)
plt.plot(epochs, loss, 'r', label='Training loss')
plt.plot(epochs, val_loss, 'b', label='Validation loss')
plt.title('Training and Validation Loss')
plt.xlabel('Epoch')
plt.ylabel('Loss Value')
plt.legend()
plt.subplot(1,4,2)
plt.plot(epochs, acc, 'r', label="Training accuracy")
plt.plot(epochs, val_acc4, 'b', label="Validation accuracy")
plt.xlabel('Epoch')
plt.ylabel('Accuracy')
plt.legend()
plt.subplot(1,4,3)
plt.plot(epochs, val_loss1, 'b', label="Loss output 1")
plt.plot(epochs, val_loss2, 'g', label="Loss output 2")
plt.plot(epochs, val_loss3, 'y', label="Loss output 3")
plt.plot(epochs, val_loss4, 'y', label="Loss output 4")
plt.legend()
plt.subplot(1,4,4)
plt.plot(epochs, val_acc1, 'b', label="Acc output 1")
plt.plot(epochs, val_acc2, 'g', label="Acc output 2")
plt.plot(epochs, val_acc3, 'y', label="Acc output 3")
plt.plot(epochs, val_acc4, 'y', label="Acc output 4")
plt.legend()
plt.show()
Finally, we visualize some predictions on the test dataset.