所以我用Keras建立了一个LSTM分类模型。
这是我的模特的片段：
x_train = sequence.pad_sequences(procprinters, maxlen = 18, padding='post', truncating='post')
x_test = sequence.pad_sequences(procprinters, maxlen = 18, padding='post', truncating='post')
x_train = array(x_train)
print(x_train.shape)
procclass = array(printerclassifcation)
permutations = np.random.permutation(procclass.shape[0])
x_train = x_train[permutations]
y_train = procclass[permutations]
print('Build model...')
model = Sequential()
model.add(Embedding(37, 128))
model.add(LSTM(128, dropout=0.1))
model.add(Dense(1, activation='sigmoid'))
model.compile(loss = 'binary_crossentropy', optimizer = 'adam', metrics=['acc'])
model.fit(x_train,y_train, validation_split = 0.2, batch_size=5, epoch
上面的代码本质上在我的X_train and y_train的前面加了0，长度为18。然后将其传递给模型进行训练。
现在，我希望对CNTK进行同样的处理。我无法在线阅读文档。有人可以帮我弄清楚我的脚垫吗？
x_train
y_train
当我使用Keras时。任何帮助或指导表示赞赏！