Tensorflow.keras：输入的形状是（），即使形状是（768、8）

Question

I hope I am doing this right (first post).

I tried to use tensorflow.keras to classify the data from here . I am aware the input shape, shape of input data and shape of targets are important when passing arguments into tf.keras.Sequential.fit()

The message I get is:

ValueError: Error when checking input: expected dense_159_input to have 2 dimensions, but got array with shape ()

So here is what I have done:

def loadDataset(file_name):
    data = pd.read_csv("data/" + file_name)
    # print(data.head(10))

    data = data.to_numpy()
    random.seed(20)

    X = data[:, 0:8]
    y = data[:, -1]

    X = np.asarray(X).reshape(X.shape[0], X.shape[1])
    X = tf.keras.utils.normalize(X, axis=0)
    y = np.asarray(y).reshape(y.shape[0], 1)

    return X, y


 title = "datasets_228_482_diabetes.csv"

 X, y = loadDataset(title)


 print(X.shape)
 print(y.shape)

(768, 8) (768, 1)

from tensorflow.keras import Sequential
from tensorflow.keras.layers import Dense, Flatten
from tensorflow.keras.optimizers import Adam

model = Sequential()
model.add(Dense(4, activation = "relu", input_shape=(8,)))
model.add(Dense(4, activation = "relu"))
model.add(Dense(1, activation = "sigmoid"))
model.compile(optimizer = "Adam",
              loss = "binary_crossentropy",
              metrics=["accuracy"])

model.fit(X, y, batch_size = 16, epochs = 1, validation_data = 0.1)

I tried making the shapes X (768, 8, 1) and y (768, 1, 1) instead in case that was the issue but then the error says it expected 2 dimensions but got three. Which makes total sense to me. I just don't understand the error above saying that the input data X has no shape when X is of shape (768, 8).

Any help would be greatly appreciated! Cheers

Answer 1

I would say that the error comes from validation_data which is supposed to be as X , some data shaped (..., 8) . Since you are passing 0.1 the dense layer doesn't understand what you are giving him.