为什么我的多输入 Keras 模型的 tf.data.Dataset.from_generator 会产生此错误？

Question

我正在尝试制作一个多输入 Keras 模型，该模型采用形状为(None, 500, 500, 3) ，类型为tensorflow.python.framework.ops.EagerTensor ，以及具有形状的一维输入(None, 1500, 1)并且是numpy.ndarray类型。 我以前有一个看起来像这样的数据集：

train_images = []
for image_path in tqdm(train_df['paths']):
  byte_file = tf.io.read_file(image_path)
  img = tf.image.decode_png(byte_file)
  train_images.append(img) 

train_1D = tf.expand_dims(encoded_strings,axis=2)
training_dataset = tf.data.Dataset.from_tensor_slices(((train_1D, train_images), training_Labels)) 
training_dataset = training_dataset.batch(20)

这对于较小的数据集效果很好，但是当我移动到较大的数据集时，我遇到了 ram 问题。 所以我决定使用tf.data.Dataset.from_generator ：

def generator():

  for input_1D in encoded_strings:
    train_1D = tf.expand_dims(input_1D, axis=2)
  
  for image_path in train_df['paths']:
    byte_file = tf.io.read_file(image_path)
    train_images = tf.image.decode_png(byte_file)

  for training_label in training_Labels:
    label = training_label
  
  yield train_1D, train_images, label


types = ((tf.int32, tf.uint8), tf.float32)


shapes = (((None, 1500, 1), (None, 500, 500, 3)), (None,))

training_dataset = tf.data.Dataset.from_generator(generator,
                                      output_types=types,
                                      output_shapes=shapes
                                     )

当我执行model.fit时，我收到以下错误消息

Epoch 1/100
---------------------------------------------------------------------------
UnknownError                              Traceback (most recent call last)
<ipython-input-36-a93e4598d6b7> in <module>()
      8 #would need to change this see age/gender
      9 history = model.fit(training_dataset,validation_data=validation_dataset,
---> 10     epochs=100)
     11 
     12 model_folder = '/content/drive/MyDrive/Colab Notebooks/models'

6 frames
/usr/local/lib/python3.7/dist-packages/tensorflow/python/eager/execute.py in quick_execute(op_name, num_outputs, inputs, attrs, ctx, name)
     58     ctx.ensure_initialized()
     59     tensors = pywrap_tfe.TFE_Py_Execute(ctx._handle, device_name, op_name,
---> 60                                         inputs, attrs, num_outputs)
     61   except core._NotOkStatusException as e:
     62     if name is not None:

UnknownError: 2 root error(s) found.
  (0) Unknown:  InvalidArgumentError: Tried to expand dim index 2 for tensor with 1 dimensions. [Op:ExpandDims]
Traceback (most recent call last):

  File "/usr/local/lib/python3.7/dist-packages/tensorflow/python/ops/script_ops.py", line 249, in __call__
    ret = func(*args)

  File "/usr/local/lib/python3.7/dist-packages/tensorflow/python/autograph/impl/api.py", line 645, in wrapper
    return func(*args, **kwargs)

  File "/usr/local/lib/python3.7/dist-packages/tensorflow/python/data/ops/dataset_ops.py", line 961, in generator_py_func
    values = next(generator_state.get_iterator(iterator_id))

  File "<ipython-input-30-e4c0ff47f2ba>", line 11, in generator
    protein_input = tf.expand_dims(protein, axis=2)

  File "/usr/local/lib/python3.7/dist-packages/tensorflow/python/util/dispatch.py", line 206, in wrapper
    return target(*args, **kwargs)

  File "/usr/local/lib/python3.7/dist-packages/tensorflow/python/ops/array_ops.py", line 436, in expand_dims_v2
    return gen_array_ops.expand_dims(input, axis, name)

  File "/usr/local/lib/python3.7/dist-packages/tensorflow/python/ops/gen_array_ops.py", line 2289, in expand_dims
    _ops.raise_from_not_ok_status(e, name)

  File "/usr/local/lib/python3.7/dist-packages/tensorflow/python/framework/ops.py", line 6897, in raise_from_not_ok_status
    six.raise_from(core._status_to_exception(e.code, message), None)

  File "<string>", line 3, in raise_from

tensorflow.python.framework.errors_impl.InvalidArgumentError: Tried to expand dim index 2 for tensor with 1 dimensions. [Op:ExpandDims]


     [[{{node PyFunc}}]]
     [[IteratorGetNext]]
     [[IteratorGetNext/_6]]
  (1) Unknown:  InvalidArgumentError: Tried to expand dim index 2 for tensor with 1 dimensions. [Op:ExpandDims]
Traceback (most recent call last):

  File "/usr/local/lib/python3.7/dist-packages/tensorflow/python/ops/script_ops.py", line 249, in __call__
    ret = func(*args)

  File "/usr/local/lib/python3.7/dist-packages/tensorflow/python/autograph/impl/api.py", line 645, in wrapper
    return func(*args, **kwargs)

  File "/usr/local/lib/python3.7/dist-packages/tensorflow/python/data/ops/dataset_ops.py", line 961, in generator_py_func
    values = next(generator_state.get_iterator(iterator_id))

  File "<ipython-input-30-e4c0ff47f2ba>", line 11, in generator
    protein_input = tf.expand_dims(protein, axis=2)

  File "/usr/local/lib/python3.7/dist-packages/tensorflow/python/util/dispatch.py", line 206, in wrapper
    return target(*args, **kwargs)

  File "/usr/local/lib/python3.7/dist-packages/tensorflow/python/ops/array_ops.py", line 436, in expand_dims_v2
    return gen_array_ops.expand_dims(input, axis, name)

  File "/usr/local/lib/python3.7/dist-packages/tensorflow/python/ops/gen_array_ops.py", line 2289, in expand_dims
    _ops.raise_from_not_ok_status(e, name)

  File "/usr/local/lib/python3.7/dist-packages/tensorflow/python/framework/ops.py", line 6897, in raise_from_not_ok_status
    six.raise_from(core._status_to_exception(e.code, message), None)

  File "<string>", line 3, in raise_from

tensorflow.python.framework.errors_impl.InvalidArgumentError: Tried to expand dim index 2 for tensor with 1 dimensions. [Op:ExpandDims]


     [[{{node PyFunc}}]]
     [[IteratorGetNext]]
0 successful operations.
0 derived errors ignored. [Op:__inference_train_function_12322]

Function call stack:
train_function -> train_function

这是由于我如何使用tf.expand_dims而导致的错误吗？ 还是我误解了output_shapes ？ 谁能建议我如何解决这个错误？