Question

下面是我的测试代码，其中包含随机生成的数据。

import numpy as np 
import pandas as pd 
import keras
from keras_contrib.layers import CRF
from keras.models import Model, Input
from keras.layers import LSTM, Embedding, Dense, TimeDistributed, Dropout, Bidirectional
import keras as k

x_train = np.random.rand(500,20,10)
y_train = np.random.rand(500,6)
print(x_train.shape, y_train.shape)# (500, 20, 10) (500, 6)

n_tags = 6
inputs = Input(shape=(20,10))
word_embedding_size = 300
n_words = 1234
model = Embedding(input_dim=n_words, output_dim=word_embedding_size, input_length=20*10)(inputs)
model = Bidirectional(LSTM(units=word_embedding_size, 
                           return_sequences=True, 
                           dropout=0.5, 
                           recurrent_dropout=0.5, 
                           kernel_initializer=k.initializers.he_normal()))(model)
model = LSTM(units=word_embedding_size * 2, 
             return_sequences=True, 
             dropout=0.5, 
             recurrent_dropout=0.5, 
             kernel_initializer=k.initializers.he_normal())(model)
model = TimeDistributed(Dense(n_tags, activation="relu"))(model)  # previously softmax output layer

crf = CRF(n_tags)
out = crf(model)
model = Model(inputs, out)
adam = k.optimizers.Adam(lr=0.0005, beta_1=0.9, beta_2=0.999)
model.compile(optimizer=adam, loss=crf.loss_function, metrics=[crf.accuracy, 'accuracy'])
history = model.fit(x_train,y_train, batch_size=16, epochs=20, validation_split=0.2, verbose=1)

但出现错误

InvalidArgumentError: Shape must be rank 3 but is rank 2 for 'bidirectional_1/Tile' (op: 'Tile') with input shapes: [?,300,1], [2].

During handling of the above exception, another exception occurred:

似乎Embedding个参数不适合LSTM吗？

如何将2D数据拟合到Keras BiLSTM-CRF模型？

0 个答案: