如何挑选Keras自定义图层？

Question

我编写了一个由Layer类扩展的自定义图层类,然后我想挑选历史记录以供进一步分析,但是当我从文件重新加载pickle对象时,python引发了一个错误:

未知图层:注意.

那么,我该如何解决呢？

我都一直get_config,__getstate__并且__setstate__,但是失败了.我只想挑选keras历史,但不是模型,所以请不要告诉我带custom_object参数的保存模型方法.

Answer 1

出现此问题的原因是在转储历史记录时,它无法转储完整模型.因此,在加载它时,它找不到自定义类.

我注意到该keras.callbacks.History对象有一个属性model,并且它的不完整转储是导致此问题的原因.

你说:

我只想腌制keras的历史,而不是模型

以下是一种解决方法:

hist = model.fit(X, Y, ...)
hist.model = None

只需将model属性设置为None,即可成功转储和加载历史对象!

以下是MVCE:

from keras.models import Sequential
from keras.layers import Conv2D, Dense, Flatten, Layer
import keras.backend as K
import numpy as np
import pickle

# MyLayer from https://keras.io/layers/writing-your-own-keras-layers/
class MyLayer(Layer):

    def __init__(self, output_dim, **kwargs):
        self.output_dim = output_dim
        super(MyLayer, self).__init__(**kwargs)

    def build(self, input_shape):
        # Create a trainable weight variable for this layer.
        self.kernel = self.add_weight(name='kernel', 
                                      shape=(input_shape[1], self.output_dim),
                                      initializer='uniform',
                                      trainable=True)
        super(MyLayer, self).build(input_shape)  # Be sure to call this at the end

    def call(self, x):
        return K.dot(x, self.kernel)

    def compute_output_shape(self, input_shape):
        return (input_shape[0], self.output_dim)

model = Sequential()
model.add(Conv2D(filters=32, kernel_size=(3,3), input_shape=(28,28,3), activation='sigmoid'))
model.add(Flatten())
model.add(MyLayer(10))
model.add(Dense(3, activation='softmax'))

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

model.summary()

X = np.random.randn(64, 28, 28, 3)
Y = np.random.randint(0, high=2, size=(64,1))

hist = model.fit(X, Y, batch_size=8)

hist.model = None

with open('hist.pkl', 'wb') as f:
    pickle.dump(hist, f)

with open('hist.pkl', 'rb') as f:
    hist_reloaded = pickle.load(f)

print(hist.history)
print(hist_reloaded.history)

输出:

{'acc':[0.484375],'亏损':[6.140302091836929]}

{'acc':[0.484375],'亏损':[6.140302091836929]}

PS如果想要使用自定义图层保存keras模型,这应该会有所帮助.