如何解决具有
我正在尝试使用keras实现此模型
input = Input(shape=(50,),dtype='float32')
encoder = Embedding(37202,150,weights=[pretrained_weights],input_length=50,trainable=True)(input)
bigram_branch = Conv1D(filters=100,kernel_size=2,padding='valid',activation='relu',strides=1)(encoder)
bigram_branch = GlobalMaxPooling1D()(bigram_branch)
trigram_branch = Conv1D(filters=100,kernel_size=3,strides=1)(encoder)
trigram_branch = GlobalMaxPooling1D()(trigram_branch)
fourgram_branch = Conv1D(filters=100,kernel_size=4,strides=1)(encoder)
fourgram_branch = GlobalMaxPooling1D()(fourgram_branch)
merged = concatenate([bigram_branch,trigram_branch,fourgram_branch],axis=1)
merged = Dense(256,activation='relu')(merged)
merged = Dropout(0.2)(merged)
merged = Dense(1)(merged)
output = Activation('sigmoid')(merged)
model = Model(inputs=[input],outputs=[output])
model.compile(loss='binary_crossentropy',optimizer='adam',metrics=['accuracy'])
model.summary()
当我开始训练时,出现此错误:
ValueError: Input 0 of layer global_max_pooling1d_6 is incompatible with the layer: expected ndim=3,found ndim=4. Full shape received: [32,50,149,100]
出现此警告:
Epoch 1/3
WARNING:tensorflow:Model was constructed with shape (None,50) for input Tensor("input_5:0",shape=(None,50),dtype=float32),but it was called on an input with incompatible shape (32,150).
训练集的形状是一个numpy矩阵的数组,每个矩阵的形状为(50,150) 请帮我。谢谢。
版权声明:本文内容由互联网用户自发贡献,该文观点与技术仅代表作者本人。本站仅提供信息存储空间服务,不拥有所有权,不承担相关法律责任。如发现本站有涉嫌侵权/违法违规的内容, 请发送邮件至 dio@foxmail.com 举报,一经查实,本站将立刻删除。