python - 使用 Python Keras 训练 CNN 1D 时出现 AttributeError

Question

我曾尝试构建 CNN 1D，但解释器告诉我:

AttributeError: 'ProgbarLogger' object has no attribute 'log_values'

这是代码片段:

model = Sequential()
model.add(Conv1D(200, 20, activation='relu', padding='same',input_shape=(1154,1024))
print(model.summary())
model.compile(optimizer='adam',
          loss='sparse_categorical_crossentropy',
          metrics=['accuracy'])
history=model.fit(X, y,batch_size=10, epochs=25,validation_split=0.7)

这是错误:

Layer (type)                 Output Shape              Param #   
=================================================================
conv1d_1 (Conv1D)            (None, 1154, 200)         4096200   
=================================================================
Total params: 4,096,200
Trainable params: 4,096,200
Non-trainable params: 0
_________________________________________________________________
None

Train on 0 samples, validate on 1 samples
Epoch 1/25
Traceback (most recent call last):
  File "binary_classification.py", line 59, in <module>
    history=model.fit(X, y,batch_size=10, epochs=25,validation_split=0.7)
  File "/home/isabella/.local/lib/python3.6/site-packages/keras/engine/training.py",
line 1039, in fit
    validation_steps=validation_steps)
  File "/home/isabella/.local/lib/python3.6/site-packages/keras/engine/training_arrays.py",
line 217, in fit_loop
    callbacks.on_epoch_end(epoch, epoch_logs)
  File "/home/isabella/.local/lib/python3.6/site-packages/keras/callbacks.py",
line 79, in on_epoch_end
    callback.on_epoch_end(epoch, logs)
  File "/home/isabella/.local/lib/python3.6/site-packages/keras/callbacks.py",
line 338, in on_epoch_end
    self.progbar.update(self.seen, self.log_values)
AttributeError: 'ProgbarLogger' object has no attribute 'log_values'

X shape is :(1,1154,1024),
y shape is :(1,1154, 1 )

Answer 1

如果仔细观察，您会在堆栈跟踪输出之前看到这一行:

Train on 0 samples, validate on 1 samples

没有训练数据!为什么？那是因为您已将 validation_split 设置为 0.7 所以首先 70% 的数据点位于 X(和 y) 留作验证，其余 30% 用于训练。 X 中的数据点数量可能少于 4，因此它的 30% 将小于 1，这意味着零数据点仍用于训练。使用超过 4 个数据点或删除 validation_split 参数(或降低它，以便至少保留一个样本用于训练)。