2

I'm trying to solve classification problem. I don't know why I'm getting this error:

ValueError: Input 0 of layer sequential_9 is incompatible with the layer: : expected min_ndim=4, found ndim=3. Full shape received: [None, None, None]

This is the main code:

model = createModel()
filesPath=getFilesPathWithoutSeizure(i, indexPat)
history=model.fit_generator(generate_arrays_for_training(indexPat, filesPath, end=75)##problem here
def createModel():
    input_shape=(1,11, 3840)
    model = Sequential()
    #C1
    model.add(Conv2D(16, (5, 5), strides=( 2, 2), padding='same',activation='relu',data_format= "channels_first", input_shape=input_shape))
    model.add(keras.layers.MaxPooling2D(pool_size=( 2, 2),data_format= "channels_first",  padding='same'))
    model.add(BatchNormalization())
    #C2
    model.add(Conv2D(32, ( 3, 3), strides=(1,1), padding='same',data_format= "channels_first",  activation='relu'))#incertezza se togliere padding
    model.add(keras.layers.MaxPooling2D(pool_size=(2, 2),data_format= "channels_first", padding='same'))
    model.add(BatchNormalization())
    #c3
    model.add(Conv2D(64, (3, 3), strides=( 1,1), padding='same',data_format= "channels_first",  activation='relu'))#incertezza se togliere padding
    model.add(keras.layers.MaxPooling2D(pool_size=(2, 2),data_format= "channels_first", padding='same'))
    model.add(BatchNormalization())
    model.add(Flatten())
    model.add(Dropout(0.5))
    model.add(Dense(256, activation='sigmoid'))
    model.add(Dropout(0.5))
    model.add(Dense(2, activation='softmax'))
    opt_adam = keras.optimizers.Adam(lr=0.00001, beta_1=0.9, beta_2=0.999, epsilon=1e-08, decay=0.0)
    model.compile(loss='categorical_crossentropy', optimizer=opt_adam, metrics=['accuracy'])
    
    return model

Error:

    history=model.fit_generator(generate_arrays_for_training(indexPat, filesPath, end=75), #end=75),#It take the first 75%
  File "/home/user1/.local/lib/python3.8/site-packages/tensorflow/python/util/deprecation.py", line 324, in new_func
    return func(*args, **kwargs)
  File "/home/user1/.local/lib/python3.8/site-packages/tensorflow/python/keras/engine/training.py", line 1815, in fit_generator
    return self.fit(
  File "/home/user1/.local/lib/python3.8/site-packages/tensorflow/python/keras/engine/training.py", line 108, in _method_wrapper
    return method(self, *args, **kwargs)
  File "/home/user1/.local/lib/python3.8/site-packages/tensorflow/python/keras/engine/training.py", line 1098, in fit
    tmp_logs = train_function(iterator)
  File "/home/user1/.local/lib/python3.8/site-packages/tensorflow/python/eager/def_function.py", line 780, in __call__
    result = self._call(*args, **kwds)
  File "/home/user1/.local/lib/python3.8/site-packages/tensorflow/python/eager/def_function.py", line 823, in _call
    self._initialize(args, kwds, add_initializers_to=initializers)
  File "/home/user1/.local/lib/python3.8/site-packages/tensorflow/python/eager/def_function.py", line 696, in _initialize
    self._stateful_fn._get_concrete_function_internal_garbage_collected(  # pylint: disable=protected-access
  File "/home/user1/.local/lib/python3.8/site-packages/tensorflow/python/eager/function.py", line 2855, in _get_concrete_function_internal_garbage_collected
    graph_function, _, _ = self._maybe_define_function(args, kwargs)
  File "/home/user1/.local/lib/python3.8/site-packages/tensorflow/python/eager/function.py", line 3213, in _maybe_define_function
    graph_function = self._create_graph_function(args, kwargs)
  File "/home/user1/.local/lib/python3.8/site-packages/tensorflow/python/eager/function.py", line 3065, in _create_graph_function
    func_graph_module.func_graph_from_py_func(
  File "/home/user1/.local/lib/python3.8/site-packages/tensorflow/python/framework/func_graph.py", line 986, in func_graph_from_py_func
    func_outputs = python_func(*func_args, **func_kwargs)
  File "/home/user1/.local/lib/python3.8/site-packages/tensorflow/python/eager/def_function.py", line 600, in wrapped_fn
    return weak_wrapped_fn().__wrapped__(*args, **kwds)
  File "/home/user1/.local/lib/python3.8/site-packages/tensorflow/python/framework/func_graph.py", line 973, in wrapper
    raise e.ag_error_metadata.to_exception(e)
ValueError: in user code:

    /home/user1/.local/lib/python3.8/site-packages/tensorflow/python/keras/engine/training.py:806 train_function  *
        return step_function(self, iterator)
    /home/user1/.local/lib/python3.8/site-packages/tensorflow/python/keras/engine/training.py:796 step_function  **
        outputs = model.distribute_strategy.run(run_step, args=(data,))
    /home/user1/.local/lib/python3.8/site-packages/tensorflow/python/distribute/distribute_lib.py:1211 run
        return self._extended.call_for_each_replica(fn, args=args, kwargs=kwargs)
    /home/user1/.local/lib/python3.8/site-packages/tensorflow/python/distribute/distribute_lib.py:2585 call_for_each_replica
        return self._call_for_each_replica(fn, args, kwargs)
    /home/user1/.local/lib/python3.8/site-packages/tensorflow/python/distribute/distribute_lib.py:2945 _call_for_each_replica
        return fn(*args, **kwargs)
    /home/user1/.local/lib/python3.8/site-packages/tensorflow/python/keras/engine/training.py:789 run_step  **
        outputs = model.train_step(data)
    /home/user1/.local/lib/python3.8/site-packages/tensorflow/python/keras/engine/training.py:747 train_step
        y_pred = self(x, training=True)
    /home/user1/.local/lib/python3.8/site-packages/tensorflow/python/keras/engine/base_layer.py:975 __call__
        input_spec.assert_input_compatibility(self.input_spec, inputs,
    /home/user1/.local/lib/python3.8/site-packages/tensorflow/python/keras/engine/input_spec.py:191 assert_input_compatibility
        raise ValueError('Input ' + str(input_index) + ' of layer ' +

    ValueError: Input 0 of layer sequential_9 is incompatible with the layer: : expected min_ndim=4, found ndim=3. Full shape received: [None, None, None]
Edayildiz
  • 545
  • 7
  • 16
  • What line of code triggered that exception? Is this line somewhere in your provided code? – Arty Sep 30 '20 at 12:58
  • Can you maybe put full exception? It has no beginning part (first lines are skipped). Hence we can't see what line of code triggered exception. – Arty Sep 30 '20 at 13:03
  • What is the shape of your numpy arrays generated by function `generate_arrays_for_training()`? – Arty Sep 30 '20 at 13:07
  • @Arty yes , this is the line `history=model.fit_generator(generate_arrays_for_training(indexPat, filesPath, end=75)##problem here` , shape of numpy array generated by `generate_arrays_for_training()` (1,11,3840) – Edayildiz Sep 30 '20 at 13:09
  • @Edaildiz I understood the reason, I'll write an answer now. – Arty Sep 30 '20 at 13:11
  • @Arty thanks a lot for your help , waiting for the solution . – Edayildiz Sep 30 '20 at 13:12
  • Done! [here](https://stackoverflow.com/a/64138518/941531) is the answer. – Arty Sep 30 '20 at 13:21
  • Basically if you're training on data with batches of size 1 you should expand dims there. If your training data `X` is of numpy shape `(1, 11, 3840)` then you have to do `X = np.expand_dims(X, 0)` with [this function](https://numpy.org/doc/stable/reference/generated/numpy.expand_dims.html). – Arty Sep 30 '20 at 13:26
  • If your images for all layers should be 3-dimensional, with 1 channel then you have to expand dims of your generated training data, do `X = np.expand_dims(X, 0)` using [this function](https://numpy.org/doc/stable/reference/generated/numpy.expand_dims.html) so that your training X data is of shape `(1, 1, 11, 3840)`, e.g. batch with 1 object, only then you can have `input_shape = (1, 11, 3840)`. – Arty Sep 30 '20 at 13:29
  • Also I see that you're writing everywhere `data_format= "channels_first"`, by default all functions are `channels_last`, in order not to write this everywhere you can reshape your generated by `generate_arrays_for_training()` data just once, if it is `X` of shape `(1, 1, 11, 3840)` then you do `X = X.transpose(0, 2, 3, 1)`. And you channels will become last dimension. – Arty Sep 30 '20 at 13:35
  • Also feel free to ask extra questions if something still not working. – Arty Sep 30 '20 at 13:36
  • @Arty Thanks a lot for helping me . I don't understand why i should use `X = X.transpose(0, 2, 3, 1)` ?? – Edayildiz Sep 30 '20 at 13:43
  • It was just a suggestion how to beautify your code, after transposing you don't need to write `"channels_first"` param everywhere, but if you don't want to change code then don't do any transposing. – Arty Sep 30 '20 at 13:48

1 Answers1

3

Keras actually always hides 0-th dimension, also known as batch dimension. Everywhere where you put input_shape = (A, B, C) actually batch dimension should not be mentioned there, (A, B, C) should be a shape of one object (or image in your case). For example if you say input_shape = (1,11, 3840) then it actually means that data for training or prediction should be a numpy array of shape something like (7, 1,11, 3840), i.e. training has 7 objects in the batch. So this 7 is size of batch, number of objects to train in parallel.

So if your one object (e.g. image) is of shape (11, 3840) then you have to write input_shape = (11, 3840) everywhere, without mentioning batch size.

Why is Keras hiding 0-th batch dimension? Because keras is expecting different size of batch, today you may provide 7 objects, tomorrow 9 and same network will work for both. But shape of one object which is (11, 3840) should never change and provided data for training generated by function generate_arrays_for_training() should be always of size (BatchSize, 11, 3840), where BatchSize can vary, you may generate batch of 1 or 7 or 9 objects-images each of shape (11, 3840).

If your images for all layers should be 3-dimensional, with 1 channel then you have to expand dims of your generated training data, do X = np.expand_dims(X, 0) using this function so that your training X data is of shape (1, 1, 11, 3840), e.g. batch with 1 object, only then you can have input_shape = (1, 11, 3840).

Also I see that you're writing everywhere data_format= "channels_first", by default all functions are channels_last, in order not to write this everywhere you can reshape your generated by generate_arrays_for_training() data just once, if it is X of shape (1, 1, 11, 3840) then you do X = X.transpose(0, 2, 3, 1). And you channels will become last dimension.

Transposing moves one dimension to another place. But for your case as you have just 1 channel than instead of transposing you can just reshape e.g. X of shape (1, 1, 11, 3840) can be reshaped by X = X.reshape(1, 11, 3840, 1) and it will become of shape (1, 11, 3840, 1). This is only needed if you want not to write "channels_first" everywhere, but if you don't want to beautify your code than transposing/reshaping is not needed at all!

As I remember from my past Keras somehow doesn't like dimensions of size 1, it basically tries to remove them in several different functions, i.e. if keras sees array of shape (1, 2, 1, 3, 1, 4) it almost always tries to reshape it to be (2, 3, 4). Thus np.expand_dims() gets actually ignored. Probably the only solution in that case would be to generate batch of size 2 images at least.

You may also read my long answer, although it is a bit unrelated, it may help you to understand how training/prediction in Keras works, especially you may read last paragraphs, numbered 1-12.

Update: As appeared to be the problem was solved due to help of next modifications:

  1. In data generating function it was needed to do expanding dims twice, i.e. X = np.expand_dims(np.expand_dims(X, 0), 0).

  2. In data generating function another was needed X = X.transpose(0, 2, 3, 1).

  3. In code of network input shape set to be input_shape = (11, 3840, 1).

  4. In code of network all substrings data_format = "channels_first" were removed.

Arty
  • 14,883
  • 6
  • 36
  • 69
  • Thanks a lot for the solution. But i don't understand why should i do `X = X.transpose(0, 2, 3, 1)` ?? – Edayildiz Sep 30 '20 at 13:40
  • @Edayildiz Transposing moves one dimension to another place. But yes for your case as you have just 1 channel than instead of transposing you can just reshape e.g. `X` of shape `(1, 1, 11, 3840)` can be reshaped by `X = X.reshape(1, 11, 3840, 1)` and it will become of shape `(1, 11, 3840, 1)`. This all is needed if you want not to write `"channels_first"`, but if you don't want to beautify your code than transposing/reshaping is not needed at all! – Arty Sep 30 '20 at 13:46
  • very well , so if I write "channels_last" then my input shape should be `input_shape=(1,11,3840,1)` ?? – Edayildiz Sep 30 '20 at 13:48
  • Unfortunately I get the same error after editing my code by expanding the dimension – Edayildiz Sep 30 '20 at 13:53
  • @Edayildiz I think it could be the case, at least as I remembered from my past that Keras incorrectly handles batch of size 1, it just removes always dimensions with size 1. The only possible solution for you is to generate at least batch of size 2, please generate numpy array for training of size `(2, 1, 11, 3840)`. – Arty Sep 30 '20 at 13:56
  • @Edayildiz If batch of size 2 doesn't help tell me I'll continue with other suggestions, I have more ideas. We'll fix this issue finally. – Arty Sep 30 '20 at 13:58
  • I get this error `ValueError: Input 0 of layer max_pooling2d_46 is incompatible with the layer: expected ndim=4, found ndim=5. Full shape received: [None, 2, 16, 6, 3840]` when i write `input_shape=(2,1,11,3840)` – Edayildiz Sep 30 '20 at 14:05
  • @Edayildiz Just write `input_shape=(1,11,3840)`, you need not to put `2` (batch size) in `input_shape`, in Keras you should never specify size of batch in `input_shape`, just remove `2` from there and it should help. – Arty Sep 30 '20 at 14:08
  • @Edayildiz If your training data generator creates numpy arrays of size `(2,1,11,3840)` then in your code for param `input_shape` you need to write `input_shape=(1,11,3840)`, you should always omit batch size (`2`) in your `input_shape` param. – Arty Sep 30 '20 at 14:10
  • I get the error in the post `ValueError: Input 0 of layer sequential_28 is incompatible with the layer: : expected min_ndim=4, found ndim=3. Full shape received: [None, None, None]` – Edayildiz Sep 30 '20 at 14:13
  • @Edayildiz This error happens with batch of 2 images? Then try to do next thing for generated array `X`, if it is of shape `(2, 1, 11, 3840)` then first try to transpose like I said `X = X.transpose(0, 2, 3, 1)`. If it doesn't help I have another solution. BTW, I'll try moving this comments discussion to chat. – Arty Sep 30 '20 at 14:16
  • How to write `(2,1,11,3840)` in input_shape and you said i can't write the batch size in keras ?? – Edayildiz Sep 30 '20 at 14:23
  • Let us [continue this discussion in chat](https://chat.stackoverflow.com/rooms/222315/discussion-between-arty-and-edayildiz). – Arty Sep 30 '20 at 14:25
  • @Edayildiz Lets have a discussion in [chat room](https://chat.stackoverflow.com/rooms/222315/discussion-between-arty-and-edayildiz) . – Arty Sep 30 '20 at 14:28