kurt-o-sys commented on issue #7349: time series delay output mxnet
URL: 
https://github.com/apache/incubator-mxnet/issues/7349#issuecomment-321361000
 
 
   Oh yeah, this may be important: I'm defining the model separately from the 
training data:
   
   ```
   input <- mx.symbol.Variable("data")
   
   guess.add <- mx.symbol.slice(input, name=guesses added/removed",
                                begin=0, end=7)
   ...
   game.trend <- mx.symbol.RNN(..., name="game trend", 
                           mode="gru", num.layers=1, state.size=1)
   regression <- mx.symbol.LinearRegressionOutput(lpeaks, name="final 
regression")
   ```
   
   I asked the question above before getting into the mxnet api itself. The 
question would be: do I need to unroll it first, so I can use 
mx.model.FeedForward.create to train it? Or is there a way to use the model as 
it is and feed it with a time-series dataset for training?
 
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services

Reply via email to