def hidden_output(
X: np.ndarray,
model: tf.keras.Model = None,
layer: int = -1,
input_shape: tuple = None,
batch_size: int = int(1e10)) -> np.ndarray:
"""
Return hidden layer output from a model on a batch of instances.
Parameters
----------
X
Batch of instances.
model
tf.keras.Model.
layer
Hidden layer of model to use as output. The default of -1 would refer to the softmax layer.
input_shape
Optional input layer shape.
batch_size
Batch size used for the model predictions.
Returns
-------
Model predictions using the specified hidden layer as output layer.
"""
if input_shape and not model.inputs:
inputs = Input(shape=input_shape)
model.call(inputs)
else:
inputs = model.inputs
hidden_model = Model(inputs=inputs, outputs=model.layers[layer].output)
X_hidden = predict_batch(hidden_model, X, batch_size=batch_size)
return X_hidden
def __init__(self,
model: tf.keras.Model,
layer: int = -1,
input_shape: tuple = None) -> None:
super().__init__()
if input_shape and not model.inputs:
inputs = Input(shape=input_shape)
model.call(inputs)
else:
inputs = model.inputs
self.model = Model(inputs=inputs, outputs=model.layers[layer].output)
def __init__(
self,
model: tf.keras.Model,
layer: int = -1,
input_shape: tuple = None,
flatten: bool = False
) -> None:
super().__init__()
if input_shape and not model.inputs:
inputs = Input(shape=input_shape)
model.call(inputs)
else:
inputs = model.inputs
self.model = Model(inputs=inputs, outputs=model.layers[layer].output)
self.flatten = Flatten() if flatten else tf.identity
def train_on_batch(x, y_true, model:tf.keras.Model, optimizer):
'''
按批次训练(正向传播)
:param x: 特征集
:param y_true: 标签集
:param model: 模型对象
:param optimizer: 优化器
:param norm 正则化项
:return:
'''
# 计算梯度 => 相当于计算偏导数 \partial{loss}/\partial{\theta}
with tf.GradientTape() as tape:
y_hat = model.call(x)
# loss = tf.losses.mean_squared_error(y, y_hat)
loss = tf.losses.binary_crossentropy(y_true, y_hat, from_logits=False)
loss += model.l2_norm()
# 计算梯度
grads = tape.gradient(loss, sources=model.trainable_variables)
# 自动更新参数
optimizer.apply_gradients(grads_and_vars=zip(grads, model.trainable_variables))
return y_hat, loss
