def GrConv2D(x,
out_channel,
kernel_shape,
padding='SAME',
stride=1,
dilation_rate=1,
W_init=None,
b_init=None,
nl=tf.identity,
split=1,
use_bias=True,
data_format='channels_last'):
if data_format == 'NHWC' or data_format == 'channels_last':
data_format = 'channels_last'
elif data_format == 'NCHW' or data_format == 'channels_first':
data_format = 'channels_first'
else:
print "aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa unknown data format"
in_shape = x.get_shape().as_list()
channel_axis = 3 if data_format == 'NHWC' else 1
in_channel = in_shape[channel_axis]
assert in_channel is not None, "[GrConv2D] Input cannot have unknown channel!"
assert in_channel % split == 0
assert out_channel % split == 0
kernel_shape = shape2d(kernel_shape)
padding = padding.upper()
filter_shape = kernel_shape + [in_channel / split, out_channel]
stride = shape2d(stride)
if W_init is None:
W_init = tf.contrib.layers.variance_scaling_initializer()
if b_init is None:
b_init = tf.constant_initializer()
with rename_get_variable({'kernel': 'W', 'bias': 'b'}):
layer = tf.layers.Conv2D(filters=out_channel,
kernel_size=kernel_shape,
strides=stride,
padding=padding,
data_format=data_format,
dilation_rate=dilation_rate,
activation=lambda x: nl(x, name='output'),
use_bias=use_bias,
kernel_initializer=W_init,
bias_initializer=b_init,
trainable=True)
ret = layer.apply(x, scope=tf.get_variable_scope())
ret.variables = VariableHolder(W=layer.kernel)
if use_bias:
ret.variables.b = layer.bias
return ret
def MyDepthConv(x,
kernel_shape,
channel_mult=1,
padding='SAME',
stride=1,
rate=1,
data_format='NHWC',
W_init=None,
activation=tf.identity):
in_shape = x.get_shape().as_list()
if data_format == 'NHWC':
in_channel = in_shape[3]
stride_shape = [1, stride, stride, 1]
elif data_format == 'NCHW':
in_channel = in_shape[1]
stride_shape = [1, 1, stride, stride]
out_channel = in_channel * channel_mult
if W_init is None:
W_init = kernel_initializer
kernel_shape = shape2d(kernel_shape) #[kernel_shape, kernel_shape]
filter_shape = kernel_shape + [in_channel, channel_mult]
W = tf.get_variable('DW', filter_shape, initializer=W_init)
conv = tf.nn.depthwise_conv2d(x,
W,
stride_shape,
padding=padding,
rate=[rate, rate],
data_format=data_format)
if activation is None:
return conv
else:
return activation(conv, name='output')
def __init__(self,
folder,
size=None,
train_or_valid='train',
channel=1,
resize=None,
debug=False,
shuffle=False,
hparams=None):
super(CustomDataSet, self).__init__()
self.folder = folder
self.is_train = True if train_or_valid == 'train' else False
self.channel = int(channel)
assert self.channel in [1, 3], self.channel
if resize is not None:
resize = shape2d(resize)
self.resize = resize
self.shuffle = shuffle
self.hparams = hparams
self.images = []
self.labels = []
if self.is_train:
self.imageDir = os.path.join(folder, 'train', 'images')
self.labelDir = os.path.join(folder, 'train', 'labels')
else:
self.imageDir = os.path.join(folder, 'test', 'images')
self.labelDir = os.path.join(folder, 'test', 'labels')
self.imageFiles = natsorted(glob2.glob(self.imageDir + '/*.*'))
self.labelFiles = natsorted(glob2.glob(self.labelDir + '/*.*'))
# print(self.imageFiles, self.labelFiles)
self._size = min(size, len(self.imageFiles))
print(self._size)
def __init__(self,
folder,
types=14,
is_train='train',
channel=1,
resize=None,
debug=False,
shuffle=False,
pathology=None,
fname='train.csv',
balancing=None):
self.version = "1.0.0"
self.description = "Vinmec is a large dataset of chest X-rays\n",
self.citation = "\n"
self.folder = folder
self.types = types
self.is_train = is_train
self.channel = int(channel)
assert self.channel in [1, 3], self.channel
if self.channel == 1:
self.imread_mode = cv2.IMREAD_GRAYSCALE
else:
self.imread_mode = cv2.IMREAD_COLOR
if resize is not None:
resize = shape2d(resize)
self.resize = resize
self.debug = debug
self.shuffle = shuffle
self.csvfile = os.path.join(self.folder, fname)
print(self.folder)
# Read the csv
self.df = pd.read_csv(self.csvfile)
self.df.columns = self.df.columns.str.replace(' ', '_')
self.df = self.df.infer_objects()
self.pathology = pathology
self.balancing = balancing
if self.balancing == 'up':
self.df_majority = self.df[self.df[self.pathology] == 0]
self.df_minority = self.df[self.df[self.pathology] == 1]
print(self.df_majority[self.pathology].value_counts())
self.df_minority_upsampled = resample(
self.df_minority,
replace=True, # sample with replacement
n_samples=self.df_majority[self.pathology].value_counts()
[0], # to match majority class
random_state=123) # reproducible results
self.df_upsampled = pd.concat(
[self.df_majority, self.df_minority_upsampled])
self.df = self.df_upsampled
def __init__(self,
folder,
types=14,
is_train='train',
channel=1,
resize=None,
debug=False,
shuffle=False,
pathology=None,
fname='train.csv'):
"""[summary]
[description
Arguments:
folder {[type]} -- [description]
Keyword Arguments:
types {number} -- [description] (default: {14})
is_train {str} -- [description] (default: {'train'}, {'valid'} or {'test'})
channel {number} -- [description] (default: {1})
resize {[type]} -- [description] (default: {None})
debug {bool} -- [description] (default: {False})
shuffle {bool} -- [description] (default: {False})
fname {str} -- [description] (default: {"train.csv"})
"""
self.version = "1.0.0"
self.description = "Vinmec is a large dataset of chest X-rays\n",
self.citation = "\n"
self.folder = folder
self.types = types
self.is_train = is_train
self.channel = int(channel)
assert self.channel in [1, 3], self.channel
if self.channel == 1:
self.imread_mode = cv2.IMREAD_GRAYSCALE
else:
self.imread_mode = cv2.IMREAD_COLOR
if resize is not None:
resize = shape2d(resize)
self.resize = resize
self.debug = debug
self.shuffle = shuffle
self.csvfile = os.path.join(self.folder, fname)
print(self.folder)
# Read the csv
self.df = pd.read_csv(self.csvfile)
self.df.columns = self.df.columns.str.replace(' ', '_')
print(self.df.info())
self.pathology = pathology
def __init__(self, dir, channel=3, resize=None, shuffle=False):
"""
Args:
files (list): list of file paths.
channel (int): 1 or 3. Will convert grayscale to RGB images if channel==3.
Will produce (h, w, 1) array if channel==1.
resize (tuple): int or (h, w) tuple. If given, resize the image.
"""
self.files = [os.path.join(dir, file) for file in os.listdir(dir)]
self.channel = int(channel)
assert self.channel in [1, 3], self.channel
self.imread_mode = cv2.IMREAD_GRAYSCALE if self.channel == 1 else cv2.IMREAD_COLOR
if resize is not None:
resize = shape2d(resize)
self.resize = resize
self.shuffle = shuffle
def DynamicConvFilter(inputs,
filters,
out_channel,
kernel_shape,
stride=1,
padding='SAME'):
""" see "Dynamic Filter Networks" (NIPS 2016)
by Bert De Brabandere*, Xu Jia*, Tinne Tuytelaars and Luc Van Gool
Remarks:
This is the convolution version of a dynamic filter.
Args:
inputs : unfiltered input [b, h, w, 1] only grayscale images.
filters : learned filters of [b, k, k, 1] (dynamically generated by the network).
out_channel (int): number of output channel.
kernel_shape: (h, w) tuple or a int.
stride: (h, w) tuple or a int.
padding (str): 'valid' or 'same'. Case insensitive.
Returns
tf.Tensor named ``output``.
"""
# tf.unstack only works with known batch_size :-(
batch_size, h, w, in_channel = inputs.get_shape().as_list()
stride = shape4d(stride)
inputs = tf.unstack(inputs)
filters = tf.reshape(filters, [batch_size] + shape2d(kernel_shape) +
[in_channel, out_channel])
filters = tf.unstack(filters)
# this is ok as TF uses the cuda stream context
rsl = [
tf.nn.conv2d(
tf.reshape(d, [1, h, w, in_channel]),
tf.reshape(k,
[kernel_shape, kernel_shape, in_channel, out_channel]),
stride,
padding="SAME") for d, k in zip(inputs, filters)
]
rsl = tf.concat(rsl, axis=0, name='output')
return rsl
def __init__(self,
folder,
size=None,
train_or_valid='train',
channel=1,
resize=None,
debug=False,
shuffle=False):
super(PretrainedSNEMI, self).__init__()
self.folder = folder
self.is_train = True if train_or_valid == 'train' else False
self.channel = int(channel)
assert self.channel in [1, 3], self.channel
if resize is not None:
resize = shape2d(resize)
self.resize = resize
self.shuffle = shuffle
self._size = size
self.images = []
self.labels = []
if self.is_train:
self.imageDir = os.path.join(folder, 'trainA')
self.labelDir = os.path.join(folder, 'trainB')
else:
self.imageDir = os.path.join(folder, 'validA')
self.labelDir = os.path.join(folder, 'validB')
imageFiles = natsorted(glob2.glob(self.imageDir + '/*.*'))
labelFiles = natsorted(glob2.glob(self.labelDir + '/*.*'))
print(imageFiles, labelFiles)
for imageFile in imageFiles:
image = skimage.io.imread(imageFile)
self.images.append(image)
for labelFile in labelFiles:
label = skimage.io.imread(labelFile)
self.labels.append(label)
self.images = np.concatenate(self.images, axis=0)
self.labels = np.concatenate(self.labels, axis=0)
def __init__(self, image_ids, channel=3, resize=None, shuffle=False):
"""
Args:
files (list): list of file paths.
channel (int): 1 or 3. Will convert grayscale to RGB images if channel==3.
Will produce (h, w, 1) array if channel==1.
resize (tuple): int or (h, w) tuple. If given, resize the image.
"""
assert len(image_ids), "No image files given to ImageFromFile!"
self.image_ids = image_ids
self.channel = int(channel)
assert self.channel in [1, 3], self.channel
self.imread_mode = cv2.IMREAD_GRAYSCALE if self.channel == 1 else cv2.IMREAD_COLOR
if resize is not None:
resize = shape2d(resize)
self.resize = resize
self.shuffle = shuffle
self.mask_file_path = TRAIN_SHIP_SEGMANTATION_PATH
self.images_path = TRAIN_DIR_PATH
self.mask_renderer = MaskRenderer(masks_file_path=self.mask_file_path)
self.pad = 63
def DynamicConvFilter(inputs, filters, out_channel,
kernel_shape,
stride=1,
padding='SAME'):
""" see "Dynamic Filter Networks" (NIPS 2016)
by Bert De Brabandere*, Xu Jia*, Tinne Tuytelaars and Luc Van Gool
Remarks:
This is the convolution version of a dynamic filter.
Args:
inputs : unfiltered input [b, h, w, 1] only grayscale images.
filters : learned filters of [b, k, k, 1] (dynamically generated by the network).
out_channel (int): number of output channel.
kernel_shape: (h, w) tuple or a int.
stride: (h, w) tuple or a int.
padding (str): 'valid' or 'same'. Case insensitive.
Returns
tf.Tensor named ``output``.
"""
# tf.unstack only works with known batch_size :-(
batch_size, h, w, in_channel = inputs.get_shape().as_list()
stride = shape4d(stride)
inputs = tf.unstack(inputs)
filters = tf.reshape(filters, [batch_size] + shape2d(kernel_shape) + [in_channel, out_channel])
filters = tf.unstack(filters)
# this is ok as TF uses the cuda stream context
rsl = [tf.nn.conv2d(tf.reshape(d, [1, h, w, in_channel]),
tf.reshape(k, [kernel_shape, kernel_shape, in_channel, out_channel]),
stride, padding="SAME") for d, k in zip(inputs, filters)]
rsl = tf.concat(rsl, axis=0, name='output')
return rsl
def MyConv2D(x,
out_channel,
kernel_shape,
padding='SAME',
stride=1,
W_init=None,
b_init=None,
activation=tf.identity,
split=1,
use_bias=True,
data_format='NHWC',
wscale=True):
in_shape = x.get_shape().as_list()
channel_axis = 3 if data_format == 'NHWC' else 1
in_channel = in_shape[channel_axis]
assert in_channel is not None, "[Conv2D] Input cannot have unknown channel!"
assert in_channel % split == 0
assert out_channel % split == 0
kernel_shape = shape2d(kernel_shape)
padding = padding.upper()
filter_shape = kernel_shape + [in_channel / split, out_channel]
stride = shape4d(stride, data_format=data_format)
# see paper details -- START
fan_in = kernel_shape[0] * kernel_shape[1] * in_channel
c = tf.constant(np.sqrt(2. / fan_in), dtype=tf.float32)
# "use a trivial N(0,1)"
W_init = tf.random_normal_initializer(stddev=c)
W = tf.get_variable('W', filter_shape, initializer=W_init)
# "scale weights at runtime"
scale = tf.sqrt(tf.reduce_mean(W**2))
W = W / scale
if b_init is None:
b_init = tf.constant_initializer()
if use_bias:
b = tf.get_variable('b', [out_channel], initializer=b_init)
if split == 1:
conv = tf.nn.conv2d(x, W, stride, padding, data_format=data_format)
else:
inputs = tf.split(x, split, channel_axis)
kernels = tf.split(W, split, 3)
outputs = [
tf.nn.conv2d(i, k, stride, padding, data_format=data_format)
for i, k in zip(inputs, kernels)
]
conv = tf.concat(outputs, channel_axis)
conv = conv * scale
ret = activation(
tf.nn.bias_add(conv, b, data_format=data_format) if use_bias else conv,
name='output')
ret.variables = VariableHolder(W=W)
if use_bias:
ret.variables.b = b
return ret
def Conv2DWithTrackedMults(x,
out_channel,
kernel_shape,
network_complexity=None,
padding='SAME',
stride=1,
W_init=None,
b_init=None,
nl=tf.identity,
split=1,
use_bias=True,
data_format='NHWC'):
"""
2D convolution on 4D inputs.
Args:
x (tf.Tensor): a 4D tensor.
Must have known number of channels, but can have other unknown dimensions.
out_channel (int): number of output channel.
kernel_shape: (h, w) tuple or a int.
stride: (h, w) tuple or a int.
padding (str): 'valid' or 'same'. Case insensitive.
split (int): Split channels as used in Alexnet. Defaults to 1 (no split).
W_init: initializer for W. Defaults to `variance_scaling_initializer`.
b_init: initializer for b. Defaults to zero.
nl: a nonlinearity function.
use_bias (bool): whether to use bias.
Returns:
tf.Tensor named ``output`` with attribute `variables`.
Variable Names:
* ``W``: weights
* ``b``: bias
"""
in_shape = x.get_shape().as_list()
channel_axis = 3 if data_format == 'NHWC' else 1
in_channel = in_shape[channel_axis]
assert in_channel is not None, "[Conv2D] Input cannot have unknown channel!"
assert in_channel % split == 0
assert out_channel % split == 0
kernel_shape = shape2d(kernel_shape)
padding = padding.upper()
filter_shape = kernel_shape + [in_channel / split, out_channel]
stride = shape4d(stride, data_format=data_format)
if W_init is None:
W_init = tf.contrib.layers.variance_scaling_initializer()
if b_init is None:
b_init = tf.constant_initializer()
W = tf.get_variable('W', filter_shape, initializer=W_init)
network_complexity['weights'] += filter_shape[0] * filter_shape[
1] * filter_shape[2] * filter_shape[3]
if use_bias:
b = tf.get_variable('b', [out_channel], initializer=b_init)
network_complexity['weights'] += out_channel
assert split == 1
xsh = x.get_shape().as_list()
network_complexity['mults'] += xsh[1] * xsh[2] * filter_shape[
0] * filter_shape[1] * filter_shape[2] * filter_shape[3] / (stride[2] *
stride[2])
conv = tf.nn.conv2d(x, W, stride, padding, data_format=data_format)
ret = nl(
tf.nn.bias_add(conv, b, data_format=data_format) if use_bias else conv,
name='output')
ret.variables = VariableHolder(W=W)
if use_bias:
ret.variables.b = b
return ret
def Conv3D(
inputs,
filters,
kernel_size,
strides=(1, 1, 1),
padding='same',
data_format='channels_last',
dilation_rate=(1, 1, 1),
activation=None,
use_bias=True,
kernel_initializer=tf.contrib.layers.variance_scaling_initializer(2.0),
bias_initializer=tf.zeros_initializer(),
kernel_regularizer=None,
bias_regularizer=None,
activity_regularizer=None,
split=1):
"""
A wrapper around `tf.layers.Conv2D`.
Some differences to maintain backward-compatibility:
1. Default kernel initializer is variance_scaling_initializer(2.0).
2. Default padding is 'same'.
3. Support 'split' argument to do group conv.
Variable Names:
* ``W``: weights
* ``b``: bias
"""
if split == 1:
with rename_get_variable({'kernel': 'W', 'bias': 'b'}):
layer = tf.layers.Conv3D(filters,
kernel_size,
strides=strides,
padding=padding,
data_format=data_format,
dilation_rate=dilation_rate,
activation=activation,
use_bias=use_bias,
kernel_initializer=kernel_initializer,
bias_initializer=bias_initializer,
kernel_regularizer=kernel_regularizer,
bias_regularizer=bias_regularizer,
activity_regularizer=activity_regularizer)
ret = layer.apply(inputs, scope=tf.get_variable_scope())
ret = tf.identity(ret, name='output')
ret.variables = VariableHolder(W=layer.kernel)
if use_bias:
ret.variables.b = layer.bias
else:
# group conv implementation
data_format = get_data_format(data_format, tfmode=False)
in_shape = inputs.get_shape().as_list()
channel_axis = 3 if data_format == 'NHWC' else 1
in_channel = in_shape[channel_axis]
assert in_channel is not None, "[Conv3D] Input cannot have unknown channel!"
assert in_channel % split == 0
assert kernel_regularizer is None and bias_regularizer is None and activity_regularizer is None, \
"Not supported by group conv now!"
out_channel = filters
assert out_channel % split == 0
assert dilation_rate == (1, 1) or get_tf_version_number(
) >= 1.5, 'TF>=1.5 required for group dilated conv'
kernel_shape = shape2d(kernel_size)
filter_shape = kernel_shape + [in_channel / split, out_channel]
stride = shape4d(strides, data_format=data_format)
kwargs = dict(data_format=data_format)
if get_tf_version_number() >= 1.5:
kwargs['dilations'] = shape4d(dilation_rate,
data_format=data_format)
W = tf.get_variable('W', filter_shape, initializer=kernel_initializer)
if use_bias:
b = tf.get_variable('b', [out_channel],
initializer=bias_initializer)
inputs = tf.split(inputs, split, channel_axis)
kernels = tf.split(W, split, 3)
outputs = [
tf.nn.conv2d(i, k, stride, padding.upper(), **kwargs)
for i, k in zip(inputs, kernels)
]
conv = tf.concat(outputs, channel_axis)
if activation is None:
activation = tf.identity
ret = activation(tf.nn.bias_add(conv, b, data_format=data_format)
if use_bias else conv,
name='output')
ret.variables = VariableHolder(W=W)
if use_bias:
ret.variables.b = b
return ret
def AtrousConv2D(x,
out_channel,
kernel_shape,
padding='SAME',
rate=1,
W_init=None,
b_init=None,
nl=tf.identity,
use_bias=False,
data_format='NHWC'):
"""
2D AtrousConvolution on 4D inputs.
Args:
x (tf.Tensor): a 4D tensor.
Must have known number of channels, but can have other unknown dimensions.
out_channel (int): number of output channel.
kernel_shape: (h, w) tuple or a int.
stride: (h, w) tuple or a int.
rate: A positive int32, In the literature, the same parameter is sometimes called input stride or dilation.
padding (str): 'valid' or 'same'. Case insensitive.
W_init: initializer for W. Defaults to `variance_scaling_initializer`.
b_init: initializer for b. Defaults to zero.
nl: a nonlinearity function.
use_bias (bool): whether to use bias.
Returns:
tf.Tensor named ``output`` with attribute `variables`.
Variable Names:
* ``W``: weights
* ``b``: bias
"""
in_shape = x.get_shape().as_list()
channel_axis = 3 if data_format == 'NHWC' else 1
in_channel = in_shape[channel_axis]
assert in_channel is not None, "[AtrousConv2D] Input cannot have unknown channel!"
kernel_shape = shape2d(kernel_shape)
padding = padding.upper()
filter_shape = kernel_shape + [in_channel, out_channel]
if W_init is None:
W_init = tf.contrib.layers.variance_scaling_initializer()
if b_init is None:
b_init = tf.constant_initializer()
W = tf.get_variable('W', filter_shape, initializer=W_init)
if use_bias:
b = tf.get_variable('b', [out_channel], initializer=b_init)
conv = tf.nn.atrous_conv2d(x, W, rate, padding)
ret = nl(
tf.nn.bias_add(conv, b, data_format=data_format) if use_bias else conv,
name='output')
ret.variables = VariableHolder(W=W)
if use_bias:
ret.variables.b = b
return ret
def Conv(inputs,
filters,
kernel_size,
strides=(1, 1),
padding='same',
data_format='channels_last',
dilation_rate=(1, 1),
activation=None,
use_bias=True,
kernel_initializer=None,
bias_initializer=tf.zeros_initializer(),
kernel_regularizer=None,
bias_regularizer=None,
activity_regularizer=None,
split=1,
norm=False):
"""
Similar to `tf.layers.Conv2D`, but with some differences:
1. Default kernel initializer is variance_scaling_initializer(2.0).
2. Default padding is 'same'.
3. Support 'split' argument to do group convolution.
Variable Names:
* ``W``: weights
* ``b``: bias
"""
if kernel_initializer is None:
if get_tf_version_tuple() <= (1, 12):
kernel_initializer = tf.contrib.layers.variance_scaling_initializer(
2.0) # deprecated
else:
kernel_initializer = tf.keras.initializers.VarianceScaling(
2.0, distribution='untruncated_normal')
dilation_rate = shape2d(dilation_rate)
if True:
# group conv implementation
data_format = get_data_format(data_format, keras_mode=False)
in_shape = inputs.get_shape().as_list()
channel_axis = 3 if data_format == 'NHWC' else 1
in_channel = in_shape[channel_axis]
assert in_channel is not None, "[Conv2D] Input cannot have unknown channel!"
assert in_channel % split == 0
assert kernel_regularizer is None and bias_regularizer is None and activity_regularizer is None, \
"Not supported by group conv or dilated conv!"
out_channel = filters
assert out_channel % split == 0
assert dilation_rate == [1, 1] or get_tf_version_tuple() >= (
1, 5), 'TF>=1.5 required for dilated conv.'
kernel_shape = shape2d(kernel_size)
filter_shape = kernel_shape + [in_channel // split, out_channel]
stride = shape4d(strides, data_format=data_format)
kwargs = {"data_format": data_format}
if get_tf_version_tuple() >= (1, 5):
kwargs['dilations'] = shape4d(dilation_rate,
data_format=data_format)
# matching input dtype (ex. tf.float16) since the default dtype of variable if tf.float32
inputs_dtype = inputs.dtype
W = tf.get_variable('parseweigth',
filter_shape,
dtype=inputs_dtype,
initializer=kernel_initializer)
if norm:
use_bias = False
W = tf.reshape(W, kernel_shape + [4, in_channel // 4, out_channel])
W = tf.nn.softmax(W, 2)
W = tf.reshape(W, filter_shape)
#dynamics = tf.reduce_mean(inputs, 0)
#dynamics = tf.transpose(dynamics, [1,2,0])
#dynamics = tf.image.resize_images(dynamics, kernel_shape)
#dynamics = tf.expand_dims(dynamics, -1)
#W = W + 0.001 * dynamics #tf.random_normal(shape = tf.shape(W), mean = 0.0, stddev = 0.012, dtype = tf.float32)
#W = W *tf.random_uniform(shape=W.get_shape().as_list(), minval=0., maxval=2.)
if use_bias:
b = tf.get_variable('parsebias', [out_channel],
dtype=inputs_dtype,
initializer=bias_initializer)
if split == 1:
conv = tf.nn.conv2d(inputs, W, stride, padding.upper(), **kwargs)
else:
try:
conv = tf.nn.conv2d(inputs, W, stride, padding.upper(),
**kwargs)
except ValueError:
log_once(
"CUDNN group convolution support is only available with "
"https://github.com/tensorflow/tensorflow/pull/25818 . "
"Will fall back to a loop-based slow implementation instead!",
'warn')
ret = tf.nn.bias_add(conv, b,
data_format=data_format) if use_bias else conv
if activation is not None:
ret = activation(ret)
ret = tf.identity(ret, name='output')
ret.variables = VariableHolder(W=W)
if use_bias:
ret.variables.b = b
return ret
def __init__(self,
folder,
types=3,
is_train='train',
channel=1,
resize=None,
debug=False,
shuffle=False,
pathology=None,
fname='train.csv',
fold_idx=0,
n_folds=5):
"""[summary]
[description
Arguments:
folder {[type]} -- [description]
Keyword Arguments:
types {number} -- [description] (default: {14})
is_train {str} -- [description] (default: {'train'}, {'valid'} or {'test'})
channel {number} -- [description] (default: {1})
resize {[type]} -- [description] (default: {None})
debug {bool} -- [description] (default: {False})
shuffle {bool} -- [description] (default: {False})
fname {str} -- [description] (default: {"train.csv"})
"""
self.version = "1.0.0"
self.description = "KFoldCovidDataset is a large dataset of chest X-rays\n",
self.citation = "\n"
self.folder = folder
self.types = types
self.is_train = is_train
self.channel = int(channel)
assert self.channel in [1, 3], self.channel
if self.channel == 1:
self.imread_mode = cv2.IMREAD_GRAYSCALE
else:
self.imread_mode = cv2.IMREAD_COLOR
if resize is not None:
resize = shape2d(resize)
self.resize = resize
self.debug = debug
self.shuffle = shuffle
self.csvfile = os.path.join(self.folder, fname)
print(self.folder)
# Read the csv
self.df = pd.read_csv(self.csvfile)
self.df.columns = self.df.columns.str.replace(' ', '_')
print(self.df.info())
self.pathology = pathology
self.indices = range(0, 1)
if n_folds > 1:
kfs = StratifiedKFold(n_splits=n_folds,
random_state=2222,
shuffle=True)
index = 0
# print(self.df[['Covid', 'Pneumonia', 'No_Disease']])
for train_indices, valid_indices in kfs.split(range(len(self.df)),
y=self.df[['Covid'
]]):
if index != fold_idx:
index += 1
continue
self.indices = train_indices if self.is_train == 'train' else valid_indices
break
else:
self.indices = list(i for i in range(len(self.df)))
print(self.__len__())
print(self.indices)
def mpusim_conv2d(
inputs,
filters,
kernel_size,
strides=(1, 1),
padding='same',
data_format='channels_last',
dilation_rate=(1, 1),
activation=None,
use_bias=True,
kernel_initializer=None,
bias_initializer=tf.zeros_initializer(),
kernel_regularizer=None,
bias_regularizer=None,
activity_regularizer=None,
split=1,
activations_datatype_size_byte=1,
weights_datatype_size_byte=1,
results_datatype_size_byte=4,
systolic_array_height=256,
systolic_array_width=256,
activation_fifo_depth=8,
accumulator_array_height=4096,
log_file_output_dir='.',
model_name='unnamed'):
"""
Similar to `tf.layers.Conv2D`, but with some differences:
1. Default kernel initializer is variance_scaling_initializer(2.0).
2. Default padding is 'same'.
3. Support 'split' argument to do group convolution.
Variable Names:
* ``W``: weights
* ``b``: bias
"""
if kernel_initializer is None:
if get_tf_version_tuple() <= (1, 12):
kernel_initializer = tf.contrib.layers.variance_scaling_initializer(2.0)
else:
kernel_initializer = tf.keras.initializers.VarianceScaling(2.0, distribution='untruncated_normal')
dilation_rate = shape2d(dilation_rate)
# group conv implementation
data_format = get_data_format(data_format, keras_mode=False)
in_shape = inputs.get_shape().as_list()
channel_axis = 3 if data_format == 'NHWC' else 1
in_channel = in_shape[channel_axis]
assert in_channel is not None, "[mpusim_conv2d] Input cannot have unknown channel!"
assert in_channel % split == 0
assert kernel_regularizer is None and bias_regularizer is None and activity_regularizer is None, \
"Not supported by group conv or dilated conv!"
out_channel = filters
assert out_channel % split == 0
assert dilation_rate == [1, 1] or get_tf_version_tuple() >= (1, 5), 'TF>=1.5 required for dilated conv.'
kernel_shape = shape2d(kernel_size)
filter_shape = kernel_shape + [in_channel / split, out_channel]
stride = shape4d(strides, data_format=data_format)
kwargs = dict(data_format=data_format)
if get_tf_version_tuple() >= (1, 5):
kwargs['dilations'] = shape4d(dilation_rate, data_format=data_format)
W = tf.get_variable(
'W', filter_shape, initializer=kernel_initializer)
if use_bias:
b = tf.get_variable('b', [out_channel], initializer=bias_initializer)
if split == 1:
conv = mpu_sim_conv2d_lib.mpu_sim_conv2d(inputs,
W,
activations_datatype_size_byte,
weights_datatype_size_byte,
results_datatype_size_byte,
systolic_array_height,
systolic_array_width,
activation_fifo_depth,
accumulator_array_height,
log_file_output_dir,
model_name,
stride,
padding.upper(),
**kwargs)
else:
inputs = tf.split(inputs, split, channel_axis)
kernels = tf.split(W, split, 3)
outputs = [mpu_sim_conv2d_lib.mpu_sim_conv2d(input_block,
kernel_block,
activations_datatype_size_byte,
weights_datatype_size_byte,
results_datatype_size_byte,
systolic_array_height,
systolic_array_width,
activation_fifo_depth,
accumulator_array_height,
log_file_output_dir,
model_name,
stride,
padding.upper(),
**kwargs)
for input_block, kernel_block in zip(inputs, kernels)]
conv = tf.concat(outputs, channel_axis)
ret = tf.nn.bias_add(conv, b, data_format=data_format) if use_bias else conv
if activation is not None:
ret = activation(ret)
ret = tf.identity(ret, name='output')
ret.variables = VariableHolder(W=W)
if use_bias:
ret.variables.b=b
return ret
def MaskedConv2D(
inputs,
filters,
kernel_size,
strides=(1, 1),
padding='same',
data_format='channels_last',
dilation_rate=(1, 1),
activation=None,
use_bias=True,
kernel_initializer=None,
bias_initializer=tf.zeros_initializer(),
kernel_regularizer=None,
bias_regularizer=None,
activity_regularizer=None,
split=1,
masking=False):
"""
A wrapper around `tf.layers.Conv2D`.
Some differences to maintain backward-compatibility:
1. Default kernel initializer is variance_scaling_initializer(2.0).
2. Default padding is 'same'.
3. Support 'split' argument to do group conv.
Variable Names:
* ``W``: weights
* ``b``: bias
"""
if kernel_initializer is None:
if get_tf_version_tuple() <= (1, 12):
kernel_initializer = tf.contrib.layers.variance_scaling_initializer(2.0)
else:
kernel_initializer = tf.keras.initializers.VarianceScaling(2.0, distribution='untruncated_normal')
dilation_rate = shape2d(dilation_rate)
if (masking == False) and (split == 1) and (dilation_rate == [1, 1]):
# tf.layers.Conv2D has bugs with dilations (https://github.com/tensorflow/tensorflow/issues/26797)
with rename_get_variable({'kernel': 'W', 'bias': 'b'}):
layer = tf.layers.Conv2D(
filters,
kernel_size,
strides=strides,
padding=padding,
data_format=data_format,
dilation_rate=dilation_rate,
activation=activation,
use_bias=use_bias,
kernel_initializer=kernel_initializer,
bias_initializer=bias_initializer,
kernel_regularizer=kernel_regularizer,
bias_regularizer=bias_regularizer,
activity_regularizer=activity_regularizer,
_reuse=tf.get_variable_scope().reuse)
ret = layer.apply(inputs, scope=tf.get_variable_scope())
ret = tf.identity(ret, name='output')
ret.variables = VariableHolder(W=layer.kernel)
if use_bias:
ret.variables.b = layer.bias
else:
if masking == True:
assert split == 1, "Pruining group conv is not supported yet"
# group conv implementation
data_format = get_data_format(data_format, keras_mode=False)
in_shape = inputs.get_shape().as_list()
channel_axis = 3 if data_format == 'NHWC' else 1
in_channel = in_shape[channel_axis]
assert in_channel is not None, "[Conv2D] Input cannot have unknown channel!"
assert in_channel % split == 0
assert kernel_regularizer is None and bias_regularizer is None and activity_regularizer is None, \
"Not supported by group conv or dilated conv!"
out_channel = filters
assert out_channel % split == 0
assert dilation_rate == [1, 1] or get_tf_version_tuple() >= (1, 5), 'TF>=1.5 required for dilated conv.'
kernel_shape = shape2d(kernel_size)
filter_shape = kernel_shape + [in_channel / split, out_channel]
stride = shape4d(strides, data_format=data_format)
kwargs = dict(data_format=data_format)
if get_tf_version_tuple() >= (1, 5):
kwargs['dilations'] = shape4d(dilation_rate, data_format=data_format)
W = tf.get_variable(
'W', filter_shape, initializer=kernel_initializer)
if use_bias:
b = tf.get_variable('b', [out_channel], initializer=bias_initializer)
if split == 1:
if masking:
W = pruning.apply_mask(W)
conv = tf.nn.conv2d(inputs, W, stride, padding.upper(), **kwargs)
else:
conv = None
if get_tf_version_tuple() >= (1, 13):
try:
conv = tf.nn.conv2d(inputs, W, stride, padding.upper(), **kwargs)
except ValueError:
log_once("CUDNN group convolution support is only available with "
"https://github.com/tensorflow/tensorflow/pull/25818 . "
"Will fall back to a loop-based slow implementation instead!", 'warn')
if conv is None:
inputs = tf.split(inputs, split, channel_axis)
kernels = tf.split(W, split, 3)
outputs = [tf.nn.conv2d(i, k, stride, padding.upper(), **kwargs)
for i, k in zip(inputs, kernels)]
conv = tf.concat(outputs, channel_axis)
ret = tf.nn.bias_add(conv, b, data_format=data_format) if use_bias else conv
if activation is not None:
ret = activation(ret)
ret = tf.identity(ret, name='output')
ret.variables = VariableHolder(W=W)
if use_bias:
ret.variables.b = b
return ret