ValueError: Input 0 of layer "sequential" is incompatible with the layer: expected shape=(None, 224, 224, 3), found shape=(224, 224, 3)

TensorFlow General Discussion
,
1

Hello, I’m working on malaria detection and diagnosis with TensorFlow and I’ve exposed to the value error problem stated in the link below. Even though I follow this [tutorial] and doing what instructor did, I found out that error. After starting to search for error and reading through the message, I couldn’t find anything and even though it says there is error, I couldn’t find any even according to error message. (https://www.youtube.com/watch?v=IA3WxTTPXqQ)

import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt
import tensorflow_datasets as tfds
from keras.layers import Conv2D, MaxPool2D, Dense, Flatten, InputLayer
from keras.optimizers import Adam
from keras.losses import BinaryCrossentropy

dataset, dataset_info = tfds.load('malaria',
                                  with_info=True,
                                  as_supervised=True,
                                  shuffle_files=True,
                                  split=['train'])

print(dataset)
print(dataset_info)

def splits(dataset, TRAIN_RATIO, VAL_RATIO, TEST_RATIO):
    DATASET_SIZE = len(dataset)

    train_dataset = dataset.take(int(TRAIN_RATIO * DATASET_SIZE))

    val_test_dataset = dataset.skip(int(TRAIN_RATIO * DATASET_SIZE))
    val_dataset = val_test_dataset.take(int(VAL_RATIO * DATASET_SIZE))

    test_dataset = val_test_dataset.skip(int(VAL_RATIO * DATASET_SIZE))
    return train_dataset, val_dataset, test_dataset

TRAIN_RATIO = 0.6
VAL_RATIO = 0.2
TEST_RATIO = 0.2

#dataset = tf.data.Dataset.range(10)
train_dataset, val_dataset, test_dataset = splits(dataset[0],
                                                  TRAIN_RATIO,
                                                  VAL_RATIO,
                                                  TEST_RATIO)
print(list(train_dataset.take(1).as_numpy_iterator()),
      list(val_dataset.take(1).as_numpy_iterator()),
      list(test_dataset.take(1).as_numpy_iterator()))

for data in dataset[0].take(4):
    print(data)

for i, (image, label) in enumerate(train_dataset.take(16)):
    ax = plt.subplot(4, 4, i + 1)
    plt.imshow(image)
    plt.title(dataset_info.features['label'].int2str(label))
    plt.axis('off')
    plt.show()

print(dataset_info.features['label'].int2str(0))

IM_SIZE = 224

def resize_rescale(image, label):
    return tf.image.resize(image, (IM_SIZE, IM_SIZE))/255.0, label

train_dataset = train_dataset.map(resize_rescale)

for image, label in train_dataset.take(1):
    print(image, label)

train_dataset = train_dataset.shuffle(buffer_size = 8,
                                      reshuffle_each_iteration = True)\
    .prefetch(tf.data.AUTOTUNE)

lenet_model = tf.keras.Sequential([
    InputLayer(input_shape=(IM_SIZE, IM_SIZE, 3)),

    Conv2D(filters = 6, kernel_size = 5, strides=1, padding='valid',
           activation= 'sigmoid'),
    MaxPool2D(pool_size=2, strides=2),

    Conv2D(filters = 16, kernel_size = 5, strides=1, padding='valid',
           activation= 'sigmoid'),
    MaxPool2D(pool_size=2, strides=2),

    Flatten(),

    Dense(100, activation= "sigmoid"),
    Dense(10, activation= "sigmoid"),
    Dense(1, activation= "sigmoid"),
])

print(lenet_model.summary())

lenet_model.compile(optimizer=Adam(learning_rate=0.01), loss=BinaryCrossentropy())

history = lenet_model.fit(train_dataset, validation_data = val_dataset,
                    epochs = 100, verbose = 1)
1 Like
3

Hi @Ata_Tekeli, You are getting this error because you are passing only input data to the model, but the model requires batch size also.

In the above code line please add the batch size.

train_dataset = train_dataset.shuffle(buffer_size = 8,
                                      reshuffle_each_iteration = True)\
                                      .batch(32).prefetch(tf.data.AUTOTUNE)

Thank You.

4

Right now working on it and made some headway, and did some progress yesterday, I’ll ask more depends on the progress.

5

And now this is the code I worked with and has the error even though some progress is made

import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt
import tensorflow_datasets as tfds
from keras.layers import Conv2D, MaxPool2D, Dense, Flatten, InputLayer
from keras.optimizers import Adam
from keras.losses import BinaryCrossentropy

dataset, dataset_info = tfds.load('malaria',
                                  with_info=True,
                                  as_supervised=True,
                                  shuffle_files=True,
                                  split=['train'])

print(dataset)
print(dataset_info)

def splits(dataset, TRAIN_RATIO, VAL_RATIO, TEST_RATIO):
    DATASET_SIZE = len(dataset)

    train_dataset = dataset.take(int(TRAIN_RATIO * DATASET_SIZE))

    val_test_dataset = dataset.skip(int(TRAIN_RATIO * DATASET_SIZE))
    val_dataset = val_test_dataset.take(int(VAL_RATIO * DATASET_SIZE))

    test_dataset = val_test_dataset.skip(int(VAL_RATIO * DATASET_SIZE))
    return train_dataset, val_dataset, test_dataset

TRAIN_RATIO = 0.6
VAL_RATIO = 0.2
TEST_RATIO = 0.2

#dataset = tf.data.Dataset.range(10)
train_dataset, val_dataset, test_dataset = splits(dataset[0],
                                                  TRAIN_RATIO,
                                                  VAL_RATIO,
                                                  TEST_RATIO)
print(list(train_dataset.take(1).as_numpy_iterator()),
      list(val_dataset.take(1).as_numpy_iterator()),
      list(test_dataset.take(1).as_numpy_iterator()))

for data in dataset[0].take(4):
    print(data)

for i, (image, label) in enumerate(train_dataset.take(16)):
    ax = plt.subplot(4, 4, i + 1)
    plt.imshow(image)
    plt.title(dataset_info.features['label'].int2str(label))
    plt.axis('off')
    plt.show()

print(dataset_info.features['label'].int2str(0))

IM_SIZE = 224
def resize_rescale(image, label):
    return tf.image.resize(image, (IM_SIZE, IM_SIZE))/255.0, label

train_dataset = train_dataset.map(resize_rescale)
val_dataset = val_dataset.map(resize_rescale)
test_dataset = test_dataset.map(resize_rescale)
print(train_dataset)
print(val_dataset)
print(test_dataset)

for image, label in train_dataset.take(1):
    print(image, label)

BATCH_SIZE = 32
train_dataset = train_dataset.shuffle(buffer_size = 8,
                                      reshuffle_each_iteration = True)\
    .prefetch(tf.data.AUTOTUNE)

val_dataset = val_dataset.shuffle(buffer_size = 8,
                                      reshuffle_each_iteration = True)\
    .prefetch(tf.data.AUTOTUNE)

print(train_dataset)
print(val_dataset)

lenet_model = tf.keras.Sequential([
    InputLayer(input_shape=(IM_SIZE, IM_SIZE, 3)),

    Conv2D(filters = 6, kernel_size = 5, strides=1, padding='valid',
           activation= 'sigmoid'),
    MaxPool2D(pool_size=2, strides=2),

    Conv2D(filters = 16, kernel_size = 5, strides=1, padding='valid',
           activation= 'sigmoid'),
    MaxPool2D(pool_size=2, strides=2),

    Flatten(),

    Dense(100, activation= "sigmoid"),
    Dense(10, activation= "sigmoid"),
    Dense(1, activation= "sigmoid"),
])

print(lenet_model.summary())

lenet_model.compile(optimizer=Adam(learning_rate=0.01), loss=BinaryCrossentropy())

history = lenet_model.fit(train_dataset, validation_data = val_dataset,
                    epochs = 100, verbose = 1)
6

Worked like a charm, thought that I could get away with defining batch size earlier on, did on both training and validation datasets