ValueError: Error when checking input: expected vgg16_input to have shape (224, 224, 3) but got array with shape (224, 224, 1)

Question

I'm trying to create a model for facial expression recognition. For this, I'm using transfer learning.

Here is the train_generator:

IMG_HEIGHT=224 
IMG_WIDTH = 224
batch_size=32

train_generator = train_datagen.flow_from_directory(
    train_data_dir,
    color_mode='grayscale',
    target_size=(IMG_HEIGHT, IMG_WIDTH),
    batch_size=batch_size,
    class_mode='categorical',
    shuffle=True)

But in this step, I'm getting the error ValueError: Error when checking input: expected vgg16_input to have shape (224, 224, 3) but got array with shape (224, 224, 1). I set the input shapes as (224,224,3), but I'm still getting the same error. I also checked the previous questions and solutions, but they didn't work. I tried train_generator=np.array(train_generator), and there is no change. What should I do to solve this error?

epochs=50
#train_generator=np.array(train_generator)
history=model.fit(train_generator,
                steps_per_epoch=num_train_imgs//batch_size,
                epochs=epochs,
                validation_data=validation_generator,
                validation_steps=num_test_imgs//batch_size)

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-25-abee8cf0777d> in <module>
      1 epochs=50
      2 #train_generator=np.array(train_generator)
----> 3 history=model.fit(train_generator,
      4                 steps_per_epoch=num_train_imgs//batch_size,
      5                 epochs=epochs,

~\anaconda3\lib\site-packages\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)
   1131             training_utils.check_generator_arguments(
   1132                 y, sample_weight, validation_split=validation_split)
-> 1133             return self.fit_generator(
   1134                 x,
   1135                 steps_per_epoch=steps_per_epoch,

~\anaconda3\lib\site-packages\keras\legacy\interfaces.py in wrapper(*args, **kwargs)
     89                 warnings.warn('Update your `' + object_name + '` call to the ' +
     90                               'Keras 2 API: ' + signature, stacklevel=2)
---> 91             return func(*args, **kwargs)
     92         wrapper._original_function = func
     93         return wrapper

~\anaconda3\lib\site-packages\keras\engine\training.py in fit_generator(self, generator, steps_per_epoch, epochs, verbose, callbacks, validation_data, validation_steps, validation_freq, class_weight, max_queue_size, workers, use_multiprocessing, shuffle, initial_epoch)
   1716         ```
   1717         """
-> 1718         return training_generator.fit_generator(
   1719             self, generator,
   1720             steps_per_epoch=steps_per_epoch,

~\anaconda3\lib\site-packages\keras\engine\training_generator.py in fit_generator(model, generator, steps_per_epoch, epochs, verbose, callbacks, validation_data, validation_steps, validation_freq, class_weight, max_queue_size, workers, use_multiprocessing, shuffle, initial_epoch)
    215                 callbacks.on_batch_begin(batch_index, batch_logs)
    216 
--> 217                 outs = model.train_on_batch(x, y,
    218                                             sample_weight=sample_weight,
    219                                             class_weight=class_weight,

~\anaconda3\lib\site-packages\keras\engine\training.py in train_on_batch(self, x, y, sample_weight, class_weight, reset_metrics)
   1503             the display labels for the scalar outputs.
   1504         """
-> 1505         x, y, sample_weights = self._standardize_user_data(
   1506             x, y,
   1507             sample_weight=sample_weight,

~\anaconda3\lib\site-packages\keras\engine\training.py in _standardize_user_data(self, x, y, sample_weight, class_weight, check_array_lengths, batch_size)
    572 
    573         # Standardize the inputs.
--> 574         x = training_utils.standardize_input_data(
    575             x,
    576             feed_input_names,

~\anaconda3\lib\site-packages\keras\engine\training_utils.py in standardize_input_data(data, names, shapes, check_batch_axis, exception_prefix)
    139                 for dim, ref_dim in zip(data_shape, shape):
    140                     if ref_dim != dim and ref_dim:
--> 141                         raise ValueError(
    142                             'Error when checking ' + exception_prefix +
    143                             ': expected ' + names[i] + ' to have shape ' +

ValueError: Error when checking input: expected vgg16_input to have shape (224, 224, 3) but got array with shape (224, 224, 1)

Answer 1

You have color_mode='grayscale'. That is why you have 1 channel image. Set "rgb" instead.