Tensorflow Keras - AttributeError: Layer features has no inbound nodes

Question

Tensorflow 版本：1.11.0

我正在尝试将 TensorBoard 与 Tensorflow keras 模型一起用于投影仪可视化。我收到 AttributeError: Layer features has no inbound nodes。我不确定为什么在下面的简单代码中会出现此错误。我确实 google 错误，但我找不到正确的解决方案来修复它。

from os import makedirs
from os.path import exists, join
import tensorflow as tf
mnist = tf.keras.datasets.mnist

import numpy as np


(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(),
  tf.keras.layers.Dense(512, activation=tf.nn.relu),
  tf.keras.layers.Dropout(0.2),
  tf.keras.layers.Flatten(),
  tf.keras.layers.Dense(10, activation=tf.nn.relu, name='features'),
  tf.keras.layers.Dense(10, activation=tf.nn.softmax)
])
model.compile(optimizer='adam',
              loss='sparse_categorical_crossentropy',
              metrics=['accuracy'])

log_dir = "./logs"
with open(join(log_dir, 'metadata.tsv'), 'w') as f:
    np.savetxt(f, y_test)

from tensorflow.keras.callbacks import TensorBoard

tf_board_callback = TensorBoard(
                            log_dir=log_dir,
                            batch_size=32,
                            embeddings_freq=1,
                            embeddings_layer_names=['features'],
                            embeddings_metadata='metadata.tsv',
                            embeddings_data=x_test
                        )

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

Answer 1

我想你应该为顺序模型的第一层指定输入形状

Answer 2

在Keras中定义网络时，添加的第一层需要添加input_shape。

在此处查看文档：https://keras.io/getting-started/sequential-model-guide/#specifying-the-input-shape

所以对于 MNIST，你应该有类似 input_shape=(28,28,1)

这里有一个很好的例子：https://www.kaggle.com/adityaecdrid/mnist-with-keras-for-beginners-99457

Tensorflow Keras - AttributeError: Layer features has no inbound nodes

Tensorflow Keras - AttributeError: Layer features has no inbound nodes

python

keras

tensorflow

tensorboard