Question

我制作了一个模型（模型 1），它以 31 个不同的数组作为输入。它由 31 个输入层组成，在下一层（使用连接层）一起计算。这种多输入模型工作得很好。
但是，我想添加更多信息。因此，我添加了一个以一个数组作为输入的模型（模型 2），然后将两个模型（1 和 2）的结果送入一个连接层，该层作为最后一个模型（模型 3）的输入进行回归。
但是现在最后一个模型的拟合方法不起作用。

history = modelConcatenate.fit(x=[trainSet, train.drop(trait, axis=1).iloc[:, 44479:].values], 
                               y=train[trait[i]].values, 
                               validation_data=([testSet, test.drop(trait, axis=1).iloc[:, 44479:].values], 
                                                test[trait[i]].values), 
                               epochs=200, callbacks=[history], verbose=0, batch_size=10)

trainSet 是 31 个数组的数组（第一个模型的输入），train 包含模型 2 的输入。
训练时出现以下错误：

ValueError                                Traceback (most recent call last)
<ipython-input-16-a7f53b631b76> in <module>
----> 1 evaluationMulti(cnn=False)

<ipython-input-14-0cc8c088dbd1> in evaluationMulti(cnn)
     74                                          validation_data=([testSet, test.drop(trait, axis=1).iloc[:, 44479:].values], 
     75                                                           test[trait[i]].values), 
---> 76                                 epochs=200, callbacks=[history], verbose=0, batch_size=10)
     77             end = time.time()
     78             print("tempsd'entrainement: ", end-start)

/usr/local/lib/python3.7/dist-packages/tensorflow_core/python/keras/engine/training.py in fit(self, x, y, batch_size, epochs, verbose, callbacks, validation_split, validation_data, shuffle, class_weight, sample_weight, initial_epoch, steps_per_epoch, validation_steps, validation_freq, max_queue_size, workers, use_multiprocessing, **kwargs)
    725         max_queue_size=max_queue_size,
    726         workers=workers,
--> 727         use_multiprocessing=use_multiprocessing)
    728 
    729   def evaluate(self,

/usr/local/lib/python3.7/dist-packages/tensorflow_core/python/keras/engine/training_arrays.py in fit(self, model, x, y, batch_size, epochs, verbose, callbacks, validation_split, validation_data, shuffle, class_weight, sample_weight, initial_epoch, steps_per_epoch, validation_steps, validation_freq, **kwargs)
    641         steps=steps_per_epoch,
    642         validation_split=validation_split,
--> 643         shuffle=shuffle)
    644 
    645     if validation_data:

/usr/local/lib/python3.7/dist-packages/tensorflow_core/python/keras/engine/training.py in _standardize_user_data(self, x, y, sample_weight, class_weight, batch_size, check_steps, steps_name, steps, validation_split, shuffle, extract_tensors_from_dataset)
   2469           feed_input_shapes,
   2470           check_batch_axis=False,  # Don't enforce the batch size.
-> 2471           exception_prefix='input')
   2472 
   2473     # Get typespecs for the input data and sanitize it if necessary.

/usr/local/lib/python3.7/dist-packages/tensorflow_core/python/keras/engine/training_utils.py in standardize_input_data(data, names, shapes, check_batch_axis, exception_prefix)
    502   elif isinstance(data, (list, tuple)):
    503     if isinstance(data[0], (list, tuple)):
--> 504       data = [np.asarray(d) for d in data]
    505     elif len(names) == 1 and isinstance(data[0], (float, int)):
    506       data = [np.asarray(data)]

/usr/local/lib/python3.7/dist-packages/tensorflow_core/python/keras/engine/training_utils.py in <listcomp>(.0)
    502   elif isinstance(data, (list, tuple)):
    503     if isinstance(data[0], (list, tuple)):
--> 504       data = [np.asarray(d) for d in data]
    505     elif len(names) == 1 and isinstance(data[0], (float, int)):
    506       data = [np.asarray(data)]

/usr/lib/python3/dist-packages/numpy/core/numeric.py in asarray(a, dtype, order)
    536 
    537     """
--> 538     return array(a, dtype, copy=False, order=order)
    539 
    540 

ValueError: could not broadcast input array from shape (804,3482) into shape (804)

804 是训练集的大小，3482 是第一个模型的第一个数组的大小。我不明白为什么使用 31 个输入的数组适用于模型 1，但现在使用 2 个输入的数组却不起作用。

具有多个输入的 Keras 模型

0 个答案: