为数据框指定了无效的“假名”

时间:2018-07-25 09:59:07

标签: r neural-network network-analysis

我正在使用Rstudio进行一些网络分析,并使用mxnet库研究GAN以进行网络分析,然后执行以下代码:

mx.gan <-
  function(data,
           label,
           hidden_nodeG = 1,
           hidden_nodeD = 1,
           out_nodeG = 1,
           out_nodeD = 1,
           activationG = "tanh",
           activationD = "tanh",
           out_activationG = "softmax",
           out_activationD = "softmax")
  {
act <- mx.symbol.Variable("data")
m <- length(hidden_nodeG)

if (length(activationG) == 1) {
  activationG <- rep(activationG, m)
} else {
  if (length(activationG) != m) {
    stop(paste("Length of activationG should be",m))
  }
}
if (length(activationD) == 1) {
  activationD <- rep(activationD, m)
} else {
  if (length(activationD) != m) {
    stop(paste("Length of activationD should be",m))
  }
}

#GENERATOR
for (i in seq_len(m)) {
  fc <- mx.symbol.FullyConnected(act, num_hidden=hidden_nodeG[i])
  act <- mx.symbol.Activation(fc, act_type=activationG[i])

}
fc <- mx.symbol.FullyConnected(act, num_hidden=out_nodeG)
out <- switch(out_activationG,
              "rmse" = mx.symbol.LinearRegressionOutput(fc),
              "softmax" = mx.symbol.SoftmaxOutput(fc),
              "logistic" = mx.symbol.LogisticRegressionOutput(fc),
              stop("Not supported yet."))
G_sym <- mx.model.FeedForward.create(out, X=data, y=label, ctx = ctx)

#DISCRIMINATOR
act <- mx.symbol.Variable("data")
m <- length(hidden_nodeD)

for (i in seq_len(m)) {
  fc <- mx.symbol.FullyConnected(act, num_hidden=hidden_nodeD[i])
  act <- mx.symbol.Activation(fc, act_type=activationD[i])

}
fc <- mx.symbol.FullyConnected(act, num_hidden=out_nodeD)
out <- switch(out_activationD,
              "rmse" = mx.symbol.LinearRegressionOutput(fc),
              "softmax" = mx.symbol.SoftmaxOutput(fc),
              "logistic" = mx.symbol.LogisticRegressionOutput(fc),
              stop("Not supported yet."))
D_sym <- mx.model.FeedForward.create(out, X=data, y=label, ctx = ctx)

return(list(G_sym, D_sym))
}

在歧视者发表评论之前,我遇到了以下错误:

dimnames<-.data.frame*tmp*中的错误,值= list(n)):   为数据框指定了无效的“假名” 从以下位置调用:dimnames<-.data.frame*tmp*,值= list(n))“

0 个答案:

没有答案