def _fixed_layer(self,
layer_id,
prev_layers,
arc,
out_filters,
stride,
is_training,
normal_or_reduction_cell="normal"):
"""
Args:
prev_layers: cache of previous layers. for skip connections
is_training: for batch_norm
"""
assert len(prev_layers) == 2
layers = [prev_layers[0], prev_layers[1]]
layers = self._maybe_calibrate_size(layers,
out_filters,
is_training=is_training)
with tf.variable_scope("layer_base"):
x = layers[1]
inp_c = self._get_C(x)
w = create_weight("w", [1, 1, inp_c, out_filters])
x = tf.nn.relu(x)
x = tf.nn.conv2d(x,
w, [1, 1, 1, 1],
"SAME",
data_format=self.data_format)
x = batch_norm(x, is_training, data_format=self.data_format)
layers[1] = x
used = np.zeros([self.num_cells + 2], dtype=np.int32)
f_sizes = [3, 5]
for cell_id in range(self.num_cells):
with tf.variable_scope("cell_{}".format(cell_id)):
x_id = arc[4 * cell_id]
used[x_id] += 1
x_op = arc[4 * cell_id + 1]
x = layers[x_id]
x_stride = stride if x_id in [0, 1] else 1
with tf.variable_scope("x_conv"):
if x_op in [0, 1]:
f_size = f_sizes[x_op]
x = self._fixed_conv(x, f_size, out_filters, x_stride,
is_training)
elif x_op in [2, 3]:
inp_c = self._get_C(x)
if x_op == 2:
x = tf.layers.average_pooling2d(
x, [3, 3], [x_stride, x_stride],
"SAME",
data_format=self.actual_data_format)
else:
x = tf.layers.max_pooling2d(
x, [3, 3], [x_stride, x_stride],
"SAME",
data_format=self.actual_data_format)
if inp_c != out_filters:
w = create_weight("w", [1, 1, inp_c, out_filters])
x = tf.nn.relu(x)
x = tf.nn.conv2d(x,
w, [1, 1, 1, 1],
"SAME",
data_format=self.data_format)
x = batch_norm(x,
is_training,
data_format=self.data_format)
else:
inp_c = self._get_C(x)
if x_stride > 1:
assert x_stride == 2
x = self._factorized_reduction(
x, out_filters, 2, is_training)
if inp_c != out_filters:
w = create_weight("w", [1, 1, inp_c, out_filters])
x = tf.nn.relu(x)
x = tf.nn.conv2d(x,
w, [1, 1, 1, 1],
"SAME",
data_format=self.data_format)
x = batch_norm(x,
is_training,
data_format=self.data_format)
if (x_op in [0, 1, 2, 3]
and self.drop_path_keep_prob is not None
and is_training):
x = self._apply_drop_path(x, layer_id)
y_id = arc[4 * cell_id + 2]
used[y_id] += 1
y_op = arc[4 * cell_id + 3]
y = layers[y_id]
y_stride = stride if y_id in [0, 1] else 1
with tf.variable_scope("y_conv"):
if y_op in [0, 1]:
f_size = f_sizes[y_op]
y = self._fixed_conv(y, f_size, out_filters, y_stride,
is_training)
elif y_op in [2, 3]:
inp_c = self._get_C(y)
if y_op == 2:
y = tf.layers.average_pooling2d(
y, [3, 3], [y_stride, y_stride],
"SAME",
data_format=self.actual_data_format)
else:
y = tf.layers.max_pooling2d(
y, [3, 3], [y_stride, y_stride],
"SAME",
data_format=self.actual_data_format)
if inp_c != out_filters:
w = create_weight("w", [1, 1, inp_c, out_filters])
y = tf.nn.relu(y)
y = tf.nn.conv2d(y,
w, [1, 1, 1, 1],
"SAME",
data_format=self.data_format)
y = batch_norm(y,
is_training,
data_format=self.data_format)
else:
inp_c = self._get_C(y)
if y_stride > 1:
assert y_stride == 2
y = self._factorized_reduction(
y, out_filters, 2, is_training)
if inp_c != out_filters:
w = create_weight("w", [1, 1, inp_c, out_filters])
y = tf.nn.relu(y)
y = tf.nn.conv2d(y,
w, [1, 1, 1, 1],
"SAME",
data_format=self.data_format)
y = batch_norm(y,
is_training,
data_format=self.data_format)
if (y_op in [0, 1, 2, 3]
and self.drop_path_keep_prob is not None
and is_training):
y = self._apply_drop_path(y, layer_id)
out = x + y
layers.append(out)
out = self._fixed_combine(layers, used, out_filters, is_training,
normal_or_reduction_cell)
return out
def _enas_cell(self, x, curr_cell, prev_cell, op_id, out_filters):
"""Performs an enas operation specified by op_id."""
num_possible_inputs = curr_cell + 1
with tf.variable_scope("avg_pool"):
avg_pool = tf.layers.average_pooling2d(
x, [3, 3], [1, 1], "SAME", data_format=self.actual_data_format)
avg_pool_c = self._get_C(avg_pool)
if avg_pool_c != out_filters:
with tf.variable_scope("conv"):
w = create_weight(
"w", [num_possible_inputs, avg_pool_c * out_filters])
w = w[prev_cell]
w = tf.reshape(w, [1, 1, avg_pool_c, out_filters])
avg_pool = tf.nn.relu(avg_pool)
avg_pool = tf.nn.conv2d(avg_pool,
w,
strides=[1, 1, 1, 1],
padding="SAME",
data_format=self.data_format)
avg_pool = batch_norm(avg_pool,
is_training=True,
data_format=self.data_format)
with tf.variable_scope("max_pool"):
max_pool = tf.layers.max_pooling2d(
x, [3, 3], [1, 1], "SAME", data_format=self.actual_data_format)
max_pool_c = self._get_C(max_pool)
if max_pool_c != out_filters:
with tf.variable_scope("conv"):
w = create_weight(
"w", [num_possible_inputs, max_pool_c * out_filters])
w = w[prev_cell]
w = tf.reshape(w, [1, 1, max_pool_c, out_filters])
max_pool = tf.nn.relu(max_pool)
max_pool = tf.nn.conv2d(max_pool,
w,
strides=[1, 1, 1, 1],
padding="SAME",
data_format=self.data_format)
max_pool = batch_norm(max_pool,
is_training=True,
data_format=self.data_format)
x_c = self._get_C(x)
if x_c != out_filters:
with tf.variable_scope("x_conv"):
w = create_weight("w",
[num_possible_inputs, x_c * out_filters])
w = w[prev_cell]
w = tf.reshape(w, [1, 1, x_c, out_filters])
x = tf.nn.relu(x)
x = tf.nn.conv2d(x,
w,
strides=[1, 1, 1, 1],
padding="SAME",
data_format=self.data_format)
x = batch_norm(x,
is_training=True,
data_format=self.data_format)
out = [
self._enas_conv(x, curr_cell, prev_cell, 3, out_filters),
self._enas_conv(x, curr_cell, prev_cell, 5, out_filters),
avg_pool,
max_pool,
x,
]
out = tf.stack(out, axis=0)
out = out[op_id, :, :, :, :]
return out
def _factorized_reduction(self, x, out_filters, stride, is_training):
"""Reduces the shape of x without information loss due to striding."""
assert out_filters % 2 == 0, (
"Need even number of filters when using this factorized reduction."
)
if stride == 1:
with tf.variable_scope("path_conv"):
inp_c = self._get_C(x)
w = create_weight("w", [1, 1, inp_c, out_filters])
x = tf.nn.conv2d(x,
w, [1, 1, 1, 1],
"SAME",
data_format=self.data_format)
x = batch_norm(x, is_training, data_format=self.data_format)
return x
stride_spec = self._get_strides(stride)
# Skip path 1
path1 = tf.nn.avg_pool(x, [1, 1, 1, 1],
stride_spec,
"VALID",
data_format=self.data_format)
with tf.variable_scope("path1_conv"):
inp_c = self._get_C(path1)
w = create_weight("w", [1, 1, inp_c, out_filters // 2])
path1 = tf.nn.conv2d(path1,
w, [1, 1, 1, 1],
"VALID",
data_format=self.data_format)
# Skip path 2
# First pad with 0"s on the right and bottom, then shift the filter to
# include those 0"s that were added.
if self.data_format == "NHWC":
pad_arr = [[0, 0], [0, 1], [0, 1], [0, 0]]
path2 = tf.pad(x, pad_arr)[:, 1:, 1:, :]
concat_axis = 3
else:
pad_arr = [[0, 0], [0, 0], [0, 1], [0, 1]]
path2 = tf.pad(x, pad_arr)[:, :, 1:, 1:]
concat_axis = 1
path2 = tf.nn.avg_pool(path2, [1, 1, 1, 1],
stride_spec,
"VALID",
data_format=self.data_format)
with tf.variable_scope("path2_conv"):
inp_c = self._get_C(path2)
w = create_weight("w", [1, 1, inp_c, out_filters // 2])
path2 = tf.nn.conv2d(path2,
w, [1, 1, 1, 1],
"VALID",
data_format=self.data_format)
# Concat and apply BN
final_path = tf.concat(values=[path1, path2], axis=concat_axis)
final_path = batch_norm(final_path,
is_training,
data_format=self.data_format)
return final_path
def _model(self, images, is_training, reuse=False):
"""Compute the logits given the images."""
if self.fixed_arc is None:
is_training = True
with tf.variable_scope(self.name, reuse=reuse):
# the first two inputs
with tf.variable_scope("stem_conv"):
w = create_weight("w", [3, 3, 3, self.out_filters * 3])
x = tf.nn.conv2d(images,
w, [1, 1, 1, 1],
"SAME",
data_format=self.data_format)
x = batch_norm(x, is_training, data_format=self.data_format)
if self.data_format == "NHCW":
split_axis = 3
elif self.data_format == "NCHW":
split_axis = 1
else:
raise ValueError("Unknown data_format '{0}'".format(
self.data_format))
layers = [x, x]
# building layers in the micro space
out_filters = self.out_filters
for layer_id in range(self.num_layers + 2):
with tf.variable_scope("layer_{0}".format(layer_id)):
if layer_id not in self.pool_layers:
if self.fixed_arc is None:
x = self._enas_layer(layer_id, layers,
self.normal_arc, out_filters)
else:
x = self._fixed_layer(
layer_id,
layers,
self.normal_arc,
out_filters,
1,
is_training,
normal_or_reduction_cell="normal")
else:
out_filters *= 2
if self.fixed_arc is None:
x = self._factorized_reduction(
x, out_filters, 2, is_training)
layers = [layers[-1], x]
x = self._enas_layer(layer_id, layers,
self.reduce_arc, out_filters)
else:
x = self._fixed_layer(
layer_id,
layers,
self.reduce_arc,
out_filters,
2,
is_training,
normal_or_reduction_cell="reduction")
print("Layer {0:>2d}: {1}".format(layer_id, x))
layers = [layers[-1], x]
# auxiliary heads
self.num_aux_vars = 0
if (self.use_aux_heads and layer_id in self.aux_head_indices
and is_training):
print("Using aux_head at layer {0}".format(layer_id))
with tf.variable_scope("aux_head"):
aux_logits = tf.nn.relu(x)
aux_logits = tf.layers.average_pooling2d(
aux_logits, [5, 5], [3, 3],
"VALID",
data_format=self.actual_data_format)
with tf.variable_scope("proj"):
inp_c = self._get_C(aux_logits)
w = create_weight("w", [1, 1, inp_c, 128])
aux_logits = tf.nn.conv2d(
aux_logits,
w, [1, 1, 1, 1],
"SAME",
data_format=self.data_format)
aux_logits = batch_norm(
aux_logits,
is_training=True,
data_format=self.data_format)
aux_logits = tf.nn.relu(aux_logits)
with tf.variable_scope("avg_pool"):
inp_c = self._get_C(aux_logits)
hw = self._get_HW(aux_logits)
w = create_weight("w", [hw, hw, inp_c, 768])
aux_logits = tf.nn.conv2d(
aux_logits,
w, [1, 1, 1, 1],
"SAME",
data_format=self.data_format)
aux_logits = batch_norm(
aux_logits,
is_training=True,
data_format=self.data_format)
aux_logits = tf.nn.relu(aux_logits)
with tf.variable_scope("fc"):
aux_logits = global_avg_pool(
aux_logits, data_format=self.data_format)
inp_c = aux_logits.get_shape()[1].value
w = create_weight("w", [inp_c, 10])
aux_logits = tf.matmul(aux_logits, w)
self.aux_logits = aux_logits
aux_head_variables = [
var for var in tf.trainable_variables()
if (var.name.startswith(self.name)
and "aux_head" in var.name)
]
self.num_aux_vars = count_model_params(aux_head_variables)
print("Aux head uses {0} params".format(self.num_aux_vars))
x = tf.nn.relu(x)
x = global_avg_pool(x, data_format=self.data_format)
if is_training and self.keep_prob is not None and self.keep_prob < 1.0:
x = tf.nn.dropout(x, self.keep_prob)
with tf.variable_scope("fc"):
inp_c = self._get_C(x)
w = create_weight("w", [inp_c, 10])
x = tf.matmul(x, w)
return x
def _fixed_layer(self, layer_id, prev_layers, start_idx, out_filters,
is_training):
"""
Args:
layer_id: current layer
prev_layers: cache of previous layers. for skip connections
start_idx: where to start looking at. technically, we can infer this
from layer_id, but why bother...
is_training: for batch_norm
"""
inputs = prev_layers[-1]
if self.whole_channels:
if self.data_format == "NHWC":
inp_c = inputs.get_shape()[3].value
elif self.data_format == "NCHW":
inp_c = inputs.get_shape()[1].value
count = self.sample_arc[start_idx]
if count in [0, 1, 2, 3]:
size = [3, 3, 5, 5]
filter_size = size[count]
with tf.variable_scope("conv_1x1"):
w = create_weight("w", [1, 1, inp_c, out_filters])
out = tf.nn.relu(inputs)
out = tf.nn.conv2d(out,
w, [1, 1, 1, 1],
"SAME",
data_format=self.data_format)
out = batch_norm(out,
is_training,
data_format=self.data_format)
with tf.variable_scope("conv_{0}x{0}".format(filter_size)):
w = create_weight(
"w",
[filter_size, filter_size, out_filters, out_filters])
out = tf.nn.relu(out)
out = tf.nn.conv2d(out,
w, [1, 1, 1, 1],
"SAME",
data_format=self.data_format)
out = batch_norm(out,
is_training,
data_format=self.data_format)
elif count == 4:
pass
elif count == 5:
pass
else:
raise ValueError(
"Unknown operation number '{0}'".format(count))
else:
count = (
self.sample_arc[start_idx:start_idx + 2 * self.num_branches] *
self.out_filters_scale)
branches = []
total_out_channels = 0
with tf.variable_scope("branch_0"):
total_out_channels += count[1]
branches.append(
self._conv_branch(inputs, 3, is_training, count[1]))
with tf.variable_scope("branch_1"):
total_out_channels += count[3]
branches.append(
self._conv_branch(inputs,
3,
is_training,
count[3],
separable=True))
with tf.variable_scope("branch_2"):
total_out_channels += count[5]
branches.append(
self._conv_branch(inputs, 5, is_training, count[5]))
with tf.variable_scope("branch_3"):
total_out_channels += count[7]
branches.append(
self._conv_branch(inputs,
5,
is_training,
count[7],
separable=True))
if self.num_branches >= 5:
with tf.variable_scope("branch_4"):
total_out_channels += count[9]
branches.append(
self._pool_branch(inputs, is_training, count[9],
"avg"))
if self.num_branches >= 6:
with tf.variable_scope("branch_5"):
total_out_channels += count[11]
branches.append(
self._pool_branch(inputs, is_training, count[11],
"max"))
with tf.variable_scope("final_conv"):
w = create_weight("w", [1, 1, total_out_channels, out_filters])
if self.data_format == "NHWC":
branches = tf.concat(branches, axis=3)
elif self.data_format == "NCHW":
branches = tf.concat(branches, axis=1)
out = tf.nn.relu(branches)
out = tf.nn.conv2d(out,
w, [1, 1, 1, 1],
"SAME",
data_format=self.data_format)
out = batch_norm(out,
is_training,
data_format=self.data_format)
if layer_id > 0:
if self.whole_channels:
skip_start = start_idx + 1
else:
skip_start = start_idx + 2 * self.num_branches
skip = self.sample_arc[skip_start:skip_start + layer_id]
total_skip_channels = np.sum(skip) + 1
res_layers = []
for i in range(layer_id):
if skip[i] == 1:
res_layers.append(prev_layers[i])
prev = res_layers + [out]
if self.data_format == "NHWC":
prev = tf.concat(prev, axis=3)
elif self.data_format == "NCHW":
prev = tf.concat(prev, axis=1)
out = prev
with tf.variable_scope("skip"):
w = create_weight(
"w",
[1, 1, total_skip_channels * out_filters, out_filters])
out = tf.nn.relu(out)
out = tf.nn.conv2d(out,
w, [1, 1, 1, 1],
"SAME",
data_format=self.data_format)
out = batch_norm(out,
is_training,
data_format=self.data_format)
return out
def _conv_branch(self,
inputs,
filter_size,
is_training,
count,
out_filters,
ch_mul=1,
start_idx=None,
separable=False):
"""
Args:
start_idx: where to start taking the output channels. if None, assuming
fixed_arc mode
count: how many output_channels to take.
"""
if start_idx is None:
assert self.fixed_arc is not None, "you screwed up!"
if self.data_format == "NHWC":
inp_c = inputs.get_shape()[3].value
elif self.data_format == "NCHW":
inp_c = inputs.get_shape()[1].value
with tf.variable_scope("inp_conv_1"):
w = create_weight("w", [1, 1, inp_c, out_filters])
x = tf.nn.conv2d(inputs,
w, [1, 1, 1, 1],
"SAME",
data_format=self.data_format)
x = batch_norm(x, is_training, data_format=self.data_format)
x = tf.nn.relu(x)
with tf.variable_scope("out_conv_{}".format(filter_size)):
if start_idx is None:
if separable:
w_depth = create_weight("w_depth", [
self.filter_size, self.filter_size, out_filters, ch_mul
])
w_point = create_weight(
"w_point", [1, 1, out_filters * ch_mul, count])
x = tf.nn.separable_conv2d(x,
w_depth,
w_point,
strides=[1, 1, 1, 1],
padding="SAME",
data_format=self.data_format)
x = batch_norm(x,
is_training,
data_format=self.data_format)
else:
w = create_weight("w",
[filter_size, filter_size, inp_c, count])
x = tf.nn.conv2d(x,
w, [1, 1, 1, 1],
"SAME",
data_format=self.data_format)
x = batch_norm(x,
is_training,
data_format=self.data_format)
else:
if separable:
w_depth = create_weight(
"w_depth",
[filter_size, filter_size, out_filters, ch_mul])
w_point = create_weight(
"w_point", [out_filters, out_filters * ch_mul])
w_point = w_point[start_idx:start_idx + count, :]
w_point = tf.transpose(w_point, [1, 0])
w_point = tf.reshape(w_point,
[1, 1, out_filters * ch_mul, count])
x = tf.nn.separable_conv2d(x,
w_depth,
w_point,
strides=[1, 1, 1, 1],
padding="SAME",
data_format=self.data_format)
mask = tf.range(0, out_filters, dtype=tf.int32)
mask = tf.logical_and(start_idx <= mask,
mask < start_idx + count)
x = batch_norm_with_mask(x,
is_training,
mask,
out_filters,
data_format=self.data_format)
else:
w = create_weight(
"w",
[filter_size, filter_size, out_filters, out_filters])
w = tf.transpose(w, [3, 0, 1, 2])
w = w[start_idx:start_idx + count, :, :, :]
w = tf.transpose(w, [1, 2, 3, 0])
x = tf.nn.conv2d(x,
w, [1, 1, 1, 1],
"SAME",
data_format=self.data_format)
mask = tf.range(0, out_filters, dtype=tf.int32)
mask = tf.logical_and(start_idx <= mask,
mask < start_idx + count)
x = batch_norm_with_mask(x,
is_training,
mask,
out_filters,
data_format=self.data_format)
x = tf.nn.relu(x)
return x
def _enas_layer(self, layer_id, prev_layers, start_idx, out_filters,
is_training):
"""
Args:
layer_id: current layer
prev_layers: cache of previous layers. for skip connections
start_idx: where to start looking at. technically, we can infer this
from layer_id, but why bother...
is_training: for batch_norm
"""
inputs = prev_layers[-1]
if self.whole_channels:
if self.data_format == "NHWC":
inp_h = inputs.get_shape()[1].value
inp_w = inputs.get_shape()[2].value
inp_c = inputs.get_shape()[3].value
elif self.data_format == "NCHW":
inp_c = inputs.get_shape()[1].value
inp_h = inputs.get_shape()[2].value
inp_w = inputs.get_shape()[3].value
count = self.sample_arc[start_idx]
branches = {}
with tf.variable_scope("branch_0"):
y = self._conv_branch(inputs,
3,
is_training,
out_filters,
out_filters,
start_idx=0)
branches[tf.equal(count, 0)] = lambda: y
with tf.variable_scope("branch_1"):
y = self._conv_branch(inputs,
3,
is_training,
out_filters,
out_filters,
start_idx=0,
separable=True)
branches[tf.equal(count, 1)] = lambda: y
with tf.variable_scope("branch_2"):
y = self._conv_branch(inputs,
5,
is_training,
out_filters,
out_filters,
start_idx=0)
branches[tf.equal(count, 2)] = lambda: y
with tf.variable_scope("branch_3"):
y = self._conv_branch(inputs,
5,
is_training,
out_filters,
out_filters,
start_idx=0,
separable=True)
branches[tf.equal(count, 3)] = lambda: y
if self.num_branches >= 5:
with tf.variable_scope("branch_4"):
y = self._pool_branch(inputs,
is_training,
out_filters,
"avg",
start_idx=0)
branches[tf.equal(count, 4)] = lambda: y
if self.num_branches >= 6:
with tf.variable_scope("branch_5"):
y = self._pool_branch(inputs,
is_training,
out_filters,
"max",
start_idx=0)
branches[tf.equal(count, 5)] = lambda: y
if self.data_format == "NHWC":
out_shape = [self.batch_size, inp_h, inp_w, out_filters]
elif self.data_format == "NCHW":
out_shape = [self.batch_size, out_filters, inp_h, inp_w]
out = tf.case(
branches,
default=lambda: tf.constant(0, tf.float32, shape=out_shape),
exclusive=True)
else:
count = self.sample_arc[start_idx:start_idx +
2 * self.num_branches]
branches = []
with tf.variable_scope("branch_0"):
branches.append(
self._conv_branch(inputs,
3,
is_training,
count[1],
out_filters,
start_idx=count[0]))
with tf.variable_scope("branch_1"):
branches.append(
self._conv_branch(inputs,
3,
is_training,
count[3],
out_filters,
start_idx=count[2],
separable=True))
with tf.variable_scope("branch_2"):
branches.append(
self._conv_branch(inputs,
5,
is_training,
count[5],
out_filters,
start_idx=count[4]))
with tf.variable_scope("branch_3"):
branches.append(
self._conv_branch(inputs,
5,
is_training,
count[7],
out_filters,
start_idx=count[6],
separable=True))
if self.num_branches >= 5:
with tf.variable_scope("branch_4"):
branches.append(
self._pool_branch(inputs,
is_training,
count[9],
"avg",
start_idx=count[8]))
if self.num_branches >= 6:
with tf.variable_scope("branch_5"):
branches.append(
self._pool_branch(inputs,
is_training,
count[11],
"max",
start_idx=count[10]))
with tf.variable_scope("final_conv"):
w = create_weight(
"w", [self.num_branches * out_filters, out_filters])
w_mask = tf.constant(
[False] * (self.num_branches * out_filters), tf.bool)
new_range = tf.range(0,
self.num_branches * out_filters,
dtype=tf.int32)
for i in range(self.num_branches):
start = out_filters * i + count[2 * i]
new_mask = tf.logical_and(
start <= new_range,
new_range < start + count[2 * i + 1])
w_mask = tf.logical_or(w_mask, new_mask)
w = tf.boolean_mask(w, w_mask)
w = tf.reshape(w, [1, 1, -1, out_filters])
inp = prev_layers[-1]
if self.data_format == "NHWC":
branches = tf.concat(branches, axis=3)
elif self.data_format == "NCHW":
branches = tf.concat(branches, axis=1)
N = tf.shape(inp)[0]
H = inp.get_shape()[2].value
W = inp.get_shape()[3].value
branches = tf.reshape(branches, [N, -1, H, W])
out = tf.nn.conv2d(branches,
w, [1, 1, 1, 1],
"SAME",
data_format=self.data_format)
out = batch_norm(out,
is_training,
data_format=self.data_format)
out = tf.nn.relu(out)
if layer_id > 0:
if self.whole_channels:
skip_start = start_idx + 1
else:
skip_start = start_idx + 2 * self.num_branches
skip = self.sample_arc[skip_start:skip_start + layer_id]
with tf.variable_scope("skip"):
res_layers = []
for i in range(layer_id):
res_layers.append(
tf.cond(tf.equal(skip[i], 1), lambda: prev_layers[i],
lambda: tf.zeros_like(prev_layers[i])))
res_layers.append(out)
out = tf.add_n(res_layers)
out = batch_norm(out,
is_training,
data_format=self.data_format)
return out
def _enas_layer(self, layer_id, prev_layers, arc, out_filters):
"""
Args:
layer_id: current layer
prev_layers: cache of previous layers. for skip connections
start_idx: where to start looking at. technically, we can infer this
from layer_id, but why bother...
"""
assert len(prev_layers) == 2, "need exactly 2 inputs"
layers = [prev_layers[0], prev_layers[1]]
layers = self._maybe_calibrate_size(layers,
out_filters,
is_training=True)
used = []
for cell_id in range(self.num_cells):
prev_layers = tf.stack(layers, axis=0)
with tf.variable_scope("cell_{0}".format(cell_id)):
with tf.variable_scope("x"):
x_id = arc[4 * cell_id]
x_op = arc[4 * cell_id + 1]
x = prev_layers[x_id, :, :, :, :]
x = self._enas_cell(x, cell_id, x_id, x_op, out_filters)
x_used = tf.one_hot(x_id,
depth=self.num_cells + 2,
dtype=tf.int32)
with tf.variable_scope("y"):
y_id = arc[4 * cell_id + 2]
y_op = arc[4 * cell_id + 3]
y = prev_layers[y_id, :, :, :, :]
y = self._enas_cell(y, cell_id, y_id, y_op, out_filters)
y_used = tf.one_hot(y_id,
depth=self.num_cells + 2,
dtype=tf.int32)
out = x + y
used.extend([x_used, y_used])
layers.append(out)
used = tf.add_n(used)
indices = tf.where(tf.equal(used, 0))
indices = tf.to_int32(indices)
indices = tf.reshape(indices, [-1])
num_outs = tf.size(indices)
out = tf.stack(layers, axis=0)
out = tf.gather(out, indices, axis=0)
inp = prev_layers[0]
if self.data_format == "NHWC":
N = tf.shape(inp)[0]
H = tf.shape(inp)[1]
W = tf.shape(inp)[2]
C = tf.shape(inp)[3]
out = tf.transpose(out, [1, 2, 3, 0, 4])
out = tf.reshape(out, [N, H, W, num_outs * out_filters])
elif self.data_format == "NCHW":
N = tf.shape(inp)[0]
C = tf.shape(inp)[1]
H = tf.shape(inp)[2]
W = tf.shape(inp)[3]
out = tf.transpose(out, [1, 0, 2, 3, 4])
out = tf.reshape(out, [N, num_outs * out_filters, H, W])
else:
raise ValueError("Unknown data_format '{0}'".format(
self.data_format))
with tf.variable_scope("final_conv"):
w = create_weight("w",
[self.num_cells + 2, out_filters * out_filters])
w = tf.gather(w, indices, axis=0)
w = tf.reshape(w, [1, 1, num_outs * out_filters, out_filters])
out = tf.nn.relu(out)
out = tf.nn.conv2d(out,
w,
strides=[1, 1, 1, 1],
padding="SAME",
data_format=self.data_format)
out = batch_norm(out,
is_training=True,
data_format=self.data_format)
out = tf.reshape(out, tf.shape(prev_layers[0]))
return out
def _pool_branch(self,
inputs,
is_training,
count,
avg_or_max,
start_idx=None):
"""
Args:
start_idx: where to start taking the output channels. if None, assuming
fixed_arc mode
count: how many output_channels to take.
"""
if start_idx is None:
assert self.fixed_arc is not None, "you screwed up!"
if self.data_format == "NHWC":
inp_c = inputs.get_shape()[3].value
elif self.data_format == "NCHW":
inp_c = inputs.get_shape()[1].value
with tf.variable_scope("conv_1"):
x = qmodules.conv(inputs,
1,
self.out_filters,
stride=1,
padding='SAME',
data_format=self.data_format)
x = batch_norm(x, is_training, data_format=self.data_format)
x = qmodules.activation(x)
with tf.variable_scope("pool"):
if self.data_format == "NHWC":
actual_data_format = "channels_last"
elif self.data_format == "NCHW":
actual_data_format = "channels_first"
if avg_or_max == "avg":
x = qmodules.pool(x,
'AVG',
3,
1,
padding='SAME',
data_format=self.data_format)
elif avg_or_max == "max":
x = qmodules.pool(x,
'MAX',
3,
1,
padding='SAME',
data_format=self.data_format)
else:
raise ValueError("Unknown pool {}".format(avg_or_max))
if start_idx is not None:
if self.data_format == "NHWC":
x = x[:, :, :, start_idx:start_idx + count]
elif self.data_format == "NCHW":
x = x[:, start_idx:start_idx + count, :, :]
return x
def _factorized_reduction(self, x, out_filters, stride, is_training):
"""Reduces the shape of x without information loss due to striding."""
assert out_filters % 2 == 0, (
"Need even number of filters when using this factorized reduction."
)
if stride == 1:
with tf.variable_scope("path_conv"):
inp_c = self._get_C(x)
x = qmodules.conv(x,
1,
out_filters,
stride=1,
data_format=self.data_format)
x = batch_norm(x, is_training, data_format=self.data_format)
return x
stride_spec = self._get_strides(stride)
# Skip path 1
# TODO: why avg_pool with kernel 1x1?
path1 = qmodules.pool(x,
'AVG',
2,
stride,
padding='VALID',
data_format=self.data_format)
with tf.variable_scope("path1_conv"):
path1 = qmodules.conv(path1,
1,
out_filters // 2,
stride=1,
padding='SAME',
data_format=self.data_format)
# Skip path 2
# First pad with 0"s on the right and bottom, then shift the filter to
# include those 0"s that were added.
if self.data_format == "NHWC":
pad_arr = [[0, 0], [0, 1], [0, 1], [0, 0]]
path2 = tf.pad(x, pad_arr)[:, 1:, 1:, :]
concat_axis = 3
else:
pad_arr = [[0, 0], [0, 0], [0, 1], [0, 1]]
path2 = tf.pad(x, pad_arr)[:, :, 1:, 1:]
concat_axis = 1
path2 = qmodules.pool(path2,
'AVG',
1,
stride,
padding='VALID',
data_format=self.data_format)
with tf.variable_scope("path2_conv"):
path2 = qmodules.conv(path2,
1,
out_filters // 2,
stride=1,
padding='SAME',
data_format=self.data_format)
# Concat and apply BN
final_path = tf.concat(values=[path1, path2], axis=concat_axis)
final_path = batch_norm(final_path,
is_training,
data_format=self.data_format)
return final_path
def _model(self, images, is_training, reuse=False):
with tf.variable_scope(self.name, reuse=reuse):
layers = []
## compute number of ReLUs
self.channel_size = self.out_filters
self.relu_size = tf.constant(0, dtype=tf.int32)
out_filters = self.out_filters
with tf.variable_scope("stem_conv"):
w = create_weight("w", [3, 3, 3, out_filters])
x = tf.nn.conv2d(images,
w, [1, 1, 1, 1],
"SAME",
data_format=self.data_format)
x = batch_norm(x, is_training, data_format=self.data_format)
layers.append(x)
if self.whole_channels:
start_idx = 0
else:
start_idx = self.num_branches
for layer_id in range(self.num_layers):
with tf.variable_scope("layer_{0}".format(layer_id)):
if self.fixed_arc is None:
x = self._enas_layer(layer_id, layers, start_idx,
out_filters, is_training)
else:
x = self._fixed_layer(layer_id, layers, start_idx,
out_filters, is_training)
layers.append(x)
if layer_id in self.pool_layers:
self.channel_size *= 2
if self.fixed_arc is not None:
## ReLU balancing
out_filters *= 4
with tf.variable_scope("pool_at_{0}".format(layer_id)):
pooled_layers = []
for i, layer in enumerate(layers):
with tf.variable_scope("from_{0}".format(i)):
x = self._factorized_reduction(
layer, out_filters, 2, is_training)
pooled_layers.append(x)
layers = pooled_layers
if self.whole_channels:
start_idx += 1 + layer_id
else:
start_idx += 2 * self.num_branches + layer_id
print(layers[-1])
x = global_avg_pool(x, data_format=self.data_format)
if is_training:
x = tf.nn.dropout(x, self.keep_prob)
with tf.variable_scope("fc"):
if self.data_format == "NWHC":
inp_c = x.get_shape()[3].value
elif self.data_format == "NCHW":
inp_c = x.get_shape()[1].value
else:
raise ValueError("Unknown data_format {0}".format(
self.data_format))
w = create_weight("w", [inp_c, 10])
x = tf.matmul(x, w)
return x
