word2veckeras.py 文件源码

def build_keras_model_cbow(index_size,vector_size,
                           context_size,
                           #code_dim,
                           sub_batch_size=1,
                           model=None,cbow_mean=False):

    kerasmodel = Graph()
    kerasmodel.add_input(name='point' , input_shape=(sub_batch_size,), dtype='int')
    kerasmodel.add_input(name='index' , input_shape=(1,), dtype='int')
    kerasmodel.add_node(Embedding(index_size, vector_size, weights=[model.syn0]),name='embedding', input='index')
    kerasmodel.add_node(Embedding(context_size, vector_size, input_length=sub_batch_size,weights=[model.keras_syn1]),name='embedpoint', input='point')
    if cbow_mean:
        kerasmodel.add_node(Lambda(lambda x:x.mean(1),output_shape=(vector_size,)),name='average',input='embedding')
    else:
        kerasmodel.add_node(Lambda(lambda x:x.sum(1),output_shape=(vector_size,)),name='average',input='embedding')

    kerasmodel.add_node(Activation('sigmoid'), name='sigmoid',inputs=['average','embedpoint'], merge_mode='dot',dot_axes=-1)
    kerasmodel.add_output(name='code',input='sigmoid')
    kerasmodel.compile('rmsprop', {'code':'mse'})
    return kerasmodel