使用 Keras Tensorflow 2.0 获取梯度

Question

我想跟踪张量板上的梯度。 然而，由于会话中运行的语句是不是一个东西了，并tf.keras.callbacks.TensorBoard的write_grads参数depricated，我想知道如何跟踪梯度的培训期间Keras或tensorflow 2.0。

我目前的方法是为此目的创建一个新的回调类，但没有成功。 也许其他人知道如何完成这种高级的东西。

为测试创建的代码如下所示，但会独立于将梯度值打印到控制台或张量板而遇到错误。

import tensorflow as tf
from tensorflow.python.keras import backend as K

mnist = tf.keras.datasets.mnist

(x_train, y_train), (x_test, y_test) = mnist.load_data()
x_train, x_test = x_train / 255.0, x_test / 255.0

model = tf.keras.models.Sequential([
  tf.keras.layers.Flatten(input_shape=(28, 28)),
  tf.keras.layers.Dense(128, activation='relu', name='dense128'),
  tf.keras.layers.Dropout(0.2),
  tf.keras.layers.Dense(10, activation='softmax', name='dense10')
])

model.compile(optimizer='adam', loss='sparse_categorical_crossentropy', metrics=['accuracy'])


class GradientCallback(tf.keras.callbacks.Callback):
    console = True

    def on_epoch_end(self, epoch, logs=None):
        weights = [w for w in self.model.trainable_weights if 'dense' in w.name and 'bias' in w.name]
        loss = self.model.total_loss
        optimizer = self.model.optimizer
        gradients = optimizer.get_gradients(loss, weights)
        for t in gradients:
            if self.console:
                print('Tensor: {}'.format(t.name))
                print('{}\n'.format(K.get_value(t)[:10]))
            else:
                tf.summary.histogram(t.name, data=t)


file_writer = tf.summary.create_file_writer("./metrics")
file_writer.set_as_default()

# write_grads has been removed
tensorboard_cb = tf.keras.callbacks.TensorBoard(histogram_freq=1, write_grads=True)
gradient_cb = GradientCallback()

model.fit(x_train, y_train, epochs=5, callbacks=[gradient_cb, tensorboard_cb])

• 将偏差梯度打印到控制台（控制台参数 = True）导致： AttributeError: 'Tensor' object has no attribute 'numpy'
• 写入 tensorboard（控制台参数 = False）会创建： TypeError：不允许将tf.Tensor用作 Python bool 。 使用if t is not None:而不是if t:来测试是否定义了张量，并使用 TensorFlow 操作（例如 tf.cond）执行以张量值为条件的子图。

Answer 1

要计算损失对权重的梯度，请使用

with tf.GradientTape() as tape:
    loss = model(model.trainable_weights)

tape.gradient(loss, model.trainable_weights)

这在GradientTape上（可以说很差）记录在案。

我们不需要tape.watch变量，因为默认情况下会监视可训练参数。

作为一个函数，它可以写成

def gradient(model, x):
    x_tensor = tf.convert_to_tensor(x, dtype=tf.float32)
    with tf.GradientTape() as t:
        t.watch(x_tensor)
        loss = model(x_tensor)
    return t.gradient(loss, x_tensor).numpy()

Answer 2

也看看这里： https : //github.com/tensorflow/tensorflow/issues/31542#issuecomment-630495970

richardwth编写了一个 Tensorboard 子类。

我对其进行了如下调整：

class ExtendedTensorBoard(tf.keras.callbacks.TensorBoard):
    def _log_gradients(self, epoch):
        writer = self._get_writer(self._train_run_name)

        with writer.as_default(), tf.GradientTape() as g:
            # here we use test data to calculate the gradients
            features, y_true = list(val_dataset.batch(100).take(1))[0]

            y_pred = self.model(features)  # forward-propagation
            loss = self.model.compiled_loss(y_true=y_true, y_pred=y_pred)  # calculate loss
            gradients = g.gradient(loss, self.model.trainable_weights)  # back-propagation

            # In eager mode, grads does not have name, so we get names from model.trainable_weights
            for weights, grads in zip(self.model.trainable_weights, gradients):
                tf.summary.histogram(
                    weights.name.replace(':', '_') + '_grads', data=grads, step=epoch)

        writer.flush()

    def on_epoch_end(self, epoch, logs=None):
        # This function overwrites the on_epoch_end in tf.keras.callbacks.TensorBoard
        # but we do need to run the original on_epoch_end, so here we use the super function.
        super(ExtendedTensorBoard, self).on_epoch_end(epoch, logs=logs)

        if self.histogram_freq and epoch % self.histogram_freq == 0:
            self._log_gradients(epoch)