Tensorflow 和 Keras 无法加载 .ckpt 保存

因此，我使用 ModelCheckpoint 回调来保存我正在训练的模型的最佳时期。它保存时没有错误，但是当我尝试加载它时，出现错误：

2019-07-27 22:58:04.713951: W tensorflow/core/util/tensor_slice_reader.cc:95] Could not open C:\Users\Riley\PycharmProjects\myNN\cp.ckpt: Data loss: not an sstable (bad magic number): perhaps your file is in a different file format and you need to use a different restore operator?

我尝试过使用绝对/完整路径，但没有运气。我确信我可以使用 EarlyStopping，但我仍然想了解为什么会收到错误。这是我的代码：

from __future__ import absolute_import, division, print_function

import tensorflow as tf
from tensorflow import keras
import numpy as np
import matplotlib.pyplot as plt
import datetime
import statistics

(train_images, train_labels), (test_images, test_labels) = np.load("dataset.npy", allow_pickle=True)

train_images = train_images / 255
test_images = test_images / 255

train_labels = list(map(float, train_labels))
test_labels = list(map(float, test_labels))
train_labels = [i/10 for i in train_labels]
test_labels = [i/10 for i in test_labels]

'''
model = keras.Sequential([
    keras.layers.Flatten(input_shape=(128, 128)),
    keras.layers.Dense(64, activation=tf.nn.relu),
    keras.layers.Dense(1)
  ])

'''

start_time = datetime.datetime.now()

model = keras.Sequential([
    keras.layers.Conv2D(32, kernel_size=(5, 5), strides=(1, 1), activation='relu', input_shape=(128, 128, 1)),
    keras.layers.MaxPooling2D(pool_size=(2, 2), strides=(2, 2)),
    keras.layers.Dropout(0.2),
    keras.layers.Conv2D(64, (5, 5), activation='relu'),
    keras.layers.MaxPooling2D(pool_size=(2, 2)),
    keras.layers.Dropout(0.2),
    keras.layers.Flatten(),
    keras.layers.Dropout(0.5),
    keras.layers.Dense(1000, activation='relu'),
    keras.layers.Dense(1)

])

model.compile(loss='mean_absolute_error',
    optimizer=keras.optimizers.SGD(lr=0.01),
    metrics=['mean_absolute_error', 'mean_squared_error'])

train_images = train_images.reshape(328, 128, 128, 1)
test_images = test_images.reshape(82, 128, 128, 1)

model.fit(train_images, train_labels, epochs=100, callbacks=[keras.callbacks.ModelCheckpoint("cp.ckpt", monitor='mean_absolute_error', save_best_only=True, verbose=1)])

model.load_weights("cp.ckpt")

predictions = model.predict(test_images)

totalDifference = 0
for i in range(82):
    print("%s: %s" % (test_labels[i] * 10, predictions[i] * 10))
    totalDifference += abs(test_labels[i] - predictions[i])

avgDifference = totalDifference / 8.2

print("\n%s\n" % avgDifference)
print("Time Elapsed:")
print(datetime.datetime.now() - start_time)

太长了；您正在保存整个模型，同时尝试仅加载权重，这不是它的工作原理。

解释

你的模特的fit:

model.fit(
    train_images,
    train_labels,
    epochs=100,
    callbacks=[
        keras.callbacks.ModelCheckpoint(
            "cp.ckpt", monitor="mean_absolute_error", save_best_only=True, verbose=1
        )
    ],
)

As save_weights=False默认情况下ModelCheckpoint，您将整个模型保存到.ckpt.

顺便提一句。文件应命名.hdf5 or .hf5就像它一样Hierarchical Data Format 5 https://en.wikipedia.org/wiki/Hierarchical_Data_Format。由于 Windows 与扩展无关，您可能会遇到一些问题，如果tensorflow / keras依赖于该操作系统上的扩展。

另一方面，您仅加载模型的权重，而文件包含整体模型:

model.load_weights("cp.ckpt")

Tensorflow 的检查点（.cp) 机制与 Keras 的 (.hdf5），所以请注意这一点（有计划将它们更紧密地集成，请参阅here https://www.tensorflow.org/beta/guide/checkpoints and here https://www.tensorflow.org/beta/guide/saved_model).

Solution

因此，要么像现在一样使用回调，BUT use model.load("model.hdf5") or add save_weights_only=True论证ModelCheckpoint:

model.fit(
    train_images,
    train_labels,
    epochs=100,
    callbacks=[
        keras.callbacks.ModelCheckpoint(
            "weights.hdf5",
            monitor="mean_absolute_error",
            save_best_only=True,
            verbose=1,
            save_weights_only=True,  # Specify this
        )
    ],
)

你可以用你的model.load_weights("weights.hdf5").