def _resource_apply_op(self, grad, var, indices=None):
# 更新判据
cond = K.equal(self.iterations % self.grad_accum_steps, 0)
# 获取梯度
ag = self.get_slot(var, 'ag')
old_update = K.update
def new_update(x, new_x):
new_x = K.switch(cond, new_x, x)
return old_update(x, new_x)
K.update = new_update
ag_t = ag / self.grad_accum_steps
op = super(new_optimizer, self)._resource_apply_op(ag_t, var)
K.update = old_update
# 累积梯度
with tf.control_dependencies([op]):
ag_t = K.switch(cond, K.zeros_like(ag), ag)
with tf.control_dependencies([K.update(ag, ag_t)]):
if indices is None:
ag_t = K.update(ag, ag + grad)
else:
ag_t = self._resource_scatter_add(ag, indices, grad)
return ag_t
def adversarial_training(model, embedding_name, epsilon=1):
"""给模型添加对抗训练
其中model是需要添加对抗训练的keras模型，embedding_name
则是model里边Embedding层的名字。要在模型compile之后使用。
"""
if model.train_function is None: # 如果还没有训练函数
model._make_train_function() # 手动make
old_train_function = model.train_function # 备份旧的训练函数
# 查找Embedding层
for output in model.outputs:
embedding_layer = search_layer(output, embedding_name)
if embedding_layer is not None:
break
if embedding_layer is None:
raise Exception('Embedding layer not found')
# 求Embedding梯度
embeddings = embedding_layer.embeddings # Embedding矩阵
gradients = K.gradients(model.total_loss, [embeddings]) # Embedding梯度
gradients = K.zeros_like(embeddings) + gradients[0] # 转为dense tensor
# 封装为函数
inputs = (model._feed_inputs + model._feed_targets +
model._feed_sample_weights) # 所有输入层
embedding_gradients = K.function(
inputs=inputs,
outputs=[gradients],
name='embedding_gradients',
) # 封装为函数
def train_function(inputs): # 重新定义训练函数
grads = embedding_gradients(inputs)[0] # Embedding梯度
delta = epsilon * grads / (np.sqrt((grads**2).sum()) + 1e-8) # 计算扰动
K.set_value(embeddings, K.eval(embeddings) + delta) # 注入扰动
outputs = old_train_function(inputs) # 梯度下降
K.set_value(embeddings, K.eval(embeddings) - delta) # 删除扰动
return outputs
model.train_function = train_function # 覆盖原训练函数
