As I am new in Deep Learning field, I am facing a strange issue regarding the input shape of my Convolutional network (1D). The input is normalized values of 13 features and total 7866 samples are available
x_train_shape (7866, 13)
The target is 0 or 1 for each features.
y_train.shape (7866, 13)
I am using following simple 1D conv network just to see how it performs.
from keras.models import Sequential
model = Sequential()
model.add(Conv1D(filters=64, kernel_size=1,
activation='relu', input_shape=(13,1)))
model.add(Conv1D(filters=64, kernel_size=1, activation='relu'))
model.add(Dropout(0.5))
model.add(MaxPooling1D(pool_size=2))
model.add(Flatten())
model.add(Dense(100, activation='relu'))
model.add(Dense(13, activation='softmax'))
model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])
model.summary()
history = model.fit(x_train, y_train, epochs=20, verbose=0)
It shows the summary of the network but also raises this error:
Error when checking input: expected conv1d_1_input to have 3 dimensions, but got array with shape (7866, 13)
I tried to implement different solutions from here and here . So when
input_shape = (13,)
is used, it raises a different error like this:
Input 0 is incompatible with layer conv1d_1: expected ndim=3, found ndim=2
Can anyone please enlighten me what is the proper input_shape for this network and why?
NB x_train and y_train are in pandas dataframe format. Is that causing any issues, should I have to convert it to another format?
Keeping the input_shape
argument the same, you should add a dimension to your input, or " unsqueeze " it.
x_train = np.expand_dims(x_train, 0)
Its shape will now be:
(7866, 13, 1)
Reshape your training data as follows: Then run your model
x_train.reshape(7866, 13,1)
The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.