What I'm trying to do is to implement a mobile net v3 Small from scratch.
Here you can find my implementation of the mobile net:
from tensorflow.keras.layers import Conv2D, DepthwiseConv2D, Dense, GlobalAveragePooling2D
from tensorflow.keras.layers import Activation, BatchNormalization, Add, Multiply, Reshape
from tensorflow.keras import backend as K
class MobileNetBase:
def __init__(self, shape, n_class, alpha=1.0):
"""Init
# Arguments
input_shape: An integer or tuple/list of 3 integers, shape
of input tensor.
n_class: Integer, number of classes.
alpha: Integer, width multiplier.
"""
self.shape = shape
self.n_class = n_class
self.alpha = alpha
def _relu6(self, x):
"""Relu 6
"""
return K.relu(x, max_value=6.0)
def _hard_swish(self, x):
"""Hard swish
"""
return x * K.relu(x + 3.0, max_value=6.0) / 6.0
def _return_activation(self, x, nl):
"""Convolution Block
This function defines a activation choice.
# Arguments
x: Tensor, input tensor of conv layer.
nl: String, nonlinearity activation type.
# Returns
Output tensor.
"""
if nl == 'HS':
x = Activation(self._hard_swish)(x)
if nl == 'RE':
x = Activation(self._relu6)(x)
return x
def _conv_block(self, inputs, filters, kernel, strides, nl):
"""Convolution Block
This function defines a 2D convolution operation with BN and activation.
# Arguments
inputs: Tensor, input tensor of conv layer.
filters: Integer, the dimensionality of the output space.
kernel: An integer or tuple/list of 2 integers, specifying the
width and height of the 2D convolution window.
strides: An integer or tuple/list of 2 integers,
specifying the strides of the convolution along the width and height.
Can be a single integer to specify the same value for
all spatial dimensions.
nl: String, nonlinearity activation type.
# Returns
Output tensor.
"""
channel_axis = 1 if K.image_data_format() == 'channels_first' else -1
x = Conv2D(filters, kernel, padding='same', strides=strides)(inputs)
x = BatchNormalization(axis=channel_axis)(x)
return self._return_activation(x, nl)
def _squeeze(self, inputs):
"""Squeeze and Excitation.
This function defines a squeeze structure.
# Arguments
inputs: Tensor, input tensor of conv layer.
"""
input_channels = int(inputs.shape[-1])
x = GlobalAveragePooling2D()(inputs)
x = Dense(input_channels, activation='relu')(x)
x = Dense(input_channels, activation='hard_sigmoid')(x)
x = Reshape((1, 1, input_channels))(x)
x = Multiply()([inputs, x])
return x
def _bottleneck(self, inputs, filters, kernel, e, s, squeeze, nl):
"""Bottleneck
This function defines a basic bottleneck structure.
# Arguments
inputs: Tensor, input tensor of conv layer.
filters: Integer, the dimensionality of the output space.
kernel: An integer or tuple/list of 2 integers, specifying the
width and height of the 2D convolution window.
e: Integer, expansion factor.
t is always applied to the input size.
s: An integer or tuple/list of 2 integers,specifying the strides
of the convolution along the width and height.Can be a single
integer to specify the same value for all spatial dimensions.
squeeze: Boolean, Whether to use the squeeze.
nl: String, nonlinearity activation type.
# Returns
Output tensor.
"""
channel_axis = 1 if K.image_data_format() == 'channels_first' else -1
input_shape = K.int_shape(inputs)
tchannel = int(e)
cchannel = int(self.alpha * filters)
r = s == 1 and input_shape[3] == filters
x = self._conv_block(inputs, tchannel, (1, 1), (1, 1), nl)
x = DepthwiseConv2D(kernel, strides=(s, s), depth_multiplier=1, padding='same')(x)
x = BatchNormalization(axis=channel_axis)(x)
x = self._return_activation(x, nl)
if squeeze:
x = self._squeeze(x)
x = Conv2D(cchannel, (1, 1), strides=(1, 1), padding='same')(x)
x = BatchNormalization(axis=channel_axis)(x)
if r:
x = Add()([x, inputs])
return x
def build(self):
pass
#MobileNet v3 small models for Keras.
from tensorflow.keras.models import Model
from tensorflow.keras.layers import Input, Conv2D, GlobalAveragePooling2D, Reshape, Dropout
from tensorflow.keras.utils import plot_model
#from model.mobilenet_base import MobileNetBase
class MobileNetV3_Small(MobileNetBase):
def __init__(self, shape, n_class, alpha=1.0, include_top=True):
"""Init.
# Arguments
input_shape: An integer or tuple/list of 3 integers, shape
of input tensor.
n_class: Integer, number of classes.
alpha: Integer, width multiplier.
include_top: if inculde classification layer.
# Returns
MobileNetv3 model.
"""
super(MobileNetV3_Small, self).__init__(shape, n_class, alpha)
self.include_top = include_top
def build(self, plot=False):
"""build MobileNetV3 Small.
# Arguments
plot: Boolean, weather to plot model.
# Returns
model: Model, model.
"""
inputs = Input(shape=self.shape)
x = self._conv_block(inputs, 16, (3, 3), strides=(2, 2), nl='HS')
x = self._bottleneck(x, 16, (3, 3), e=16, s=2, squeeze=True, nl='RE')
x = self._bottleneck(x, 24, (3, 3), e=72, s=2, squeeze=False, nl='RE')
x = self._bottleneck(x, 24, (3, 3), e=88, s=1, squeeze=False, nl='RE')
x = self._bottleneck(x, 40, (5, 5), e=96, s=2, squeeze=True, nl='HS')
x = self._bottleneck(x, 40, (5, 5), e=240, s=1, squeeze=True, nl='HS')
x = self._bottleneck(x, 40, (5, 5), e=240, s=1, squeeze=True, nl='HS')
x = self._bottleneck(x, 48, (5, 5), e=120, s=1, squeeze=True, nl='HS')
x = self._bottleneck(x, 48, (5, 5), e=144, s=1, squeeze=True, nl='HS')
x = self._bottleneck(x, 96, (5, 5), e=288, s=2, squeeze=True, nl='HS')
x = self._bottleneck(x, 96, (5, 5), e=576, s=1, squeeze=True, nl='HS')
x = self._bottleneck(x, 96, (5, 5), e=576, s=1, squeeze=True, nl='HS')
x = self._conv_block(x, 576, (1, 1), strides=(1, 1), nl='HS')
x = GlobalAveragePooling2D()(x)
x = Reshape((1, 1, 576))(x)
x = Conv2D(1280, (1, 1), padding='same')(x)
x = self._return_activation(x, 'HS')
if self.include_top:
x = Dropout(0.5)(x)
x = Conv2D(self.n_class, (1, 1), padding='same', activation='softmax')(x)
x = Reshape((self.n_class,))(x)
model = Model(inputs, x)
if plot:
plot_model(model, to_file='images/MobileNetv3_small.png', show_shapes=True)
return model
I tried to train this net with a flowers dataset.
Here's how I loaded the dataset into the runtime:
from keras.preprocessing.image import ImageDataGenerator
train_datagen=ImageDataGenerator(rescale=1./255)
train_generator=train_datagen.flow_from_directory('D:/Grottini/Rete_mani/flower_photos_Complete/train',
target_size=(80,80),
color_mode='rgb',
batch_size=64,
class_mode='categorical',
shuffle=True)
valid_generator=train_datagen.flow_from_directory('D:/Grottini/Rete_mani/flower_photos_Complete/valid',
target_size=(80,80),
color_mode='rgb',
batch_size=64,
class_mode='categorical',
shuffle=True)
STEP_SIZE_TRAIN=train_generator.n//train_generator.batch_size
STEP_SIZE_VALID=valid_generator.n//valid_generator.batch_size
As you can see, I used a image Data generator, with a Rescale (1./255) as a preprocessing operation (as requested from mobile net) Here there are some examples of what the mobile net is going to work with:
And, finally, with these lines of code I try to train the mobile net:
history=model.fit_generator(generator=train_generator,
steps_per_epoch=STEP_SIZE_TRAIN,
validation_data=valid_generator,
validation_steps=STEP_SIZE_VALID,
epochs=20
)
But, even though I thought I was doing everything fine, my training failed, as you can see here: Training Learning Curve
Can someone please explain me what I did wrong? Or even what should I do in order to train properly this net?
