GAN LSTM Time Series

Does anyone know if it is possible to use LSTM or another RNN in GAN architecture as the generator? Here a reference: 1

Thx!

To clarify, what I am asking is it possible in MMA 12.3 (my current version) and what I have done so far:

generator = NetChain[
  {
   50,
   ParametricRampLayer[],
   50,
   ParametricRampLayer[],
   LinearLayer[{}],
   Tanh
   }
  ,
  "Output" -> "Real"
  ]

discriminator = NetChain[
 {
 50,
 ParametricRampLayer[],
 50,
 ParametricRampLayer[],
 LinearLayer[]
 }
 ,
 "Input"->"Real",
 "Output"->"Real"
 ]

gan = NetGANOperator[{generator, discriminator}, "Latent" -> 20]

This works fine, by the way it is the reference example from Wolfram GAN Net Operator Wolfram

However, I have read in some recent papers Stanford, that researchers are including in the discriminator, for example, some kind of recurrent layer. So, I have tried, without success, the following:

discriminator = NetInitialize@
 NetChain[{GatedRecurrentLayer[1], LinearLayer[1]}, "Input" -> {1, 1},
   "Output" -> "Scalar"]

But I receive the following error message:

NetGANOperator: The discriminator (second element of the first argument), should output a real number, but it outputs a length-1 vector of real numbers instead.

I have already tried to change the output format but nothing seems to work.

So, if anyone has something to add I would be very grateful.

Thank you!

EDIT after answer @Dropped Bass

    generator =  NetChain[
  {
   50,
   ParametricRampLayer[]
   ,
   50
   ,
   ParametricRampLayer[]
   ,
   LinearLayer[{10}]
   ,
   ReshapeLayer[{10, 1}]
   }
  ]

discriminator =  NetInitialize@
  NetChain[{GatedRecurrentLayer[1], LinearLayer[{}]}, 
   "Input" -> {10, 1}]

gan = NetGANOperator[{generator, discriminator}, 
  "Latent" -> {10, 1}]

It seems correct so far and here is the sample and latent function generator as is in NetGANOperator:

MatrixQ[Table[RandomVariate[NormalDistribution[], 1], {i, 1, 10}]]
Dimensions[Table[RandomVariate[NormalDistribution[], 1], {i, 1, 10}]]
sam = Table[RandomVariate[NormalDistribution[], 1], {i, 1, 10}]
MatrixQ[sam]
Dimensions[sam]
lat = Table[RandomReal[{0, 1}, 1], {i, 1, 10}]
MatrixQ[lat]
Dimensions[lat]

 NetTrain[gan,
 Function[
  <|
   "Sample" -> sam,
   "Latent" -> lat
   |>]
 ,
 TrainingUpdateSchedule -> {"Generator", "Discriminator"},
 BatchSize -> 10,
 MaxTrainingRounds -> 100,
 TargetDevice -> {"GPU",1}
 ]

But I receive this error: