華爲諾亞實驗室開源Disout算法，tensorflow 2.0 的實現。
算法通過增加擾動，而不是屏蔽某些特徵，來解決過擬合問題。
論文：https://arxiv.org/abs/2002.11022
實現代碼：
import tensorflow as tf
class Disout(tf.keras.layers.Layer):
'''
disout
論文：https://arxiv.org/abs/2002.11022
'''
def __init__(self, dist_prob, block_size=5, alpha=0.5, **kwargs):
super(Disout, self).__init__(**kwargs)
self.dist_prob = dist_prob
self.alpha = alpha
self.block_size = block_size
def build(self, input_shape):
pass
@tf.function
def call(self, x):
'''x：(batch_size,h,w,c)'''
if not self.trainable:
return x
else:
if tf.math.equal(tf.rank(x),4):
x_shape = tf.shape(x)
x_size = x_shape[1:3]
x_size_f = tf.cast(x_size, tf.float32)
# 計算block_size
x_block_size_f = tf.constant((self.block_size, self.block_size), tf.float32)
x_block_size = tf.cast(x_block_size_f, tf.int32)
# 根據dist_prob，計算block_num
x_block_num = (x_size_f[0] * x_size_f[1]) * self.dist_prob / (x_block_size_f[0] * x_block_size_f[1])
# 計算block在中心區域出現的概率
x_block_rate = x_block_num / ((x_size_f[0] - x_block_size_f[0] + 1) * (x_size_f[1] - x_block_size_f[1] + 1))
# 根據概率生成block區域
x_block_center = tf.random.normal((x_shape[0], x_size[0] - x_block_size[0] + 1, x_size[1] - x_block_size[1] + 1, x_shape[3]), dtype=tf.float32)
x_block_padding_t = x_block_size[0] // 2
x_block_padding_b = x_size_f[0] - tf.cast(x_block_padding_t, tf.float32) - (x_size_f[0] - x_block_size_f[0] + 1.0)
x_block_padding_b = tf.cast(x_block_padding_b, tf.int32)
x_block_padding_l = x_block_size[1] // 2
x_block_padding_r = x_size_f[1] - tf.cast(x_block_padding_l, tf.float32) - (x_size_f[1] - x_block_size_f[1] + 1.0)
x_block_padding_r = tf.cast(x_block_padding_r, tf.int32)
x_block_padding = tf.pad(x_block_center,[[0, 0],[x_block_padding_t, x_block_padding_b],[x_block_padding_l, x_block_padding_r],[0, 0]])
x_block = tf.cast(x_block_padding<x_block_rate, tf.float32)
x_block = tf.nn.max_pool2d(x_block, ksize=[self.block_size, self.block_size], strides=[1, 1], padding='SAME')
# 疊加擾動
x_max = tf.reduce_max(x, axis=(1,2), keepdims=True)
x_min = tf.reduce_min(x, axis=(1,2), keepdims=True)
x_block_random = tf.random.normal(x_shape, dtype=x.dtype) * (x_max - x_min) + x_min
x_block_random = x_block_random * (1.0 - self.alpha) + x * self.alpha
x = x * (1-x_block) + x_block_random * x_block
return x
else:
return x
def compute_output_shape(self, input_shape):
'''計算輸出shape'''
return input_shape
class Disout1D(tf.keras.layers.Layer):
'''
disout
論文：https://arxiv.org/abs/2002.11022
'''
def __init__(self, dist_prob, block_size=5, alpha=0.5, **kwargs):
super(Disout1D, self).__init__(**kwargs)
self.dist_prob = dist_prob
self.alpha = alpha
self.block_size = block_size
def build(self, input_shape):
pass
@tf.function
def call(self, x):
'''x：(batch_size,h,w,c)'''
if not self.trainable:
return x
else:
if tf.math.equal(tf.rank(x),2):
x_shape = tf.shape(x)
x_size = x_shape[1]
x_size_f = tf.cast(x_size, tf.float32)
# 計算block_size
x_block_size_f = tf.constant(self.block_size, tf.float32)
# x_block_size_f = x_size_f * self.block_size
# x_block_size_f = tf.math.maximum(x_block_size_f, 1)
x_block_size = tf.cast(x_block_size_f, tf.int32)
# 根據dist_prob，計算block_num
x_block_num = (x_size_f) * self.dist_prob / (x_block_size_f)
# 計算block在中心區域出現的概率
x_block_rate = x_block_num / ((x_size_f - x_block_size_f + 1))
# 根據概率生成block區域
x_block_center = tf.random.normal((x_shape[0], x_size - x_block_size + 1), dtype=tf.float32)
x_block_padding_t = x_block_size // 2
x_block_padding_b = x_size_f - tf.cast(x_block_padding_t, tf.float32) - (x_size_f - x_block_size_f + 1.0)
x_block_padding_b = tf.cast(x_block_padding_b, tf.int32)
x_block_padding = tf.pad(x_block_center,[[0, 0],[x_block_padding_t, x_block_padding_b]])
x_block = tf.cast(x_block_padding<x_block_rate, tf.float32)
x_block = tf.expand_dims(x_block, axis=-1)
x_block = tf.nn.max_pool1d(x_block, ksize=[self.block_size], strides=[1], padding='SAME')
x_block = tf.reshape(x_block, x_shape)
# 疊加擾動
x_max = tf.reduce_max(x, axis=1, keepdims=True)
x_min = tf.reduce_min(x, axis=1, keepdims=True)
x_block_random = tf.random.normal(x_shape, dtype=x.dtype) * (x_max - x_min) + x_min
x_block_random = x_block_random * (1.0 - self.alpha) + x * self.alpha
x = x * (1-x_block) + x_block_random * x_block
return x
else:
return x
def compute_output_shape(self, input_shape):
'''計算輸出shape'''
return input_shape
源碼：https://github.com/tfwcn/tensorflow2-disout
相關資料：
華爲源碼：https://github.com/huawei-noah/Disout