How to input flattened layer into LSTM or GRU

[CD]

I have a working model with some convolutional (et al) layers. I am flattening (Flatten layer) and feeding to a fully connected (Dense) layer. I would now like to feed the flattened layer into an LSTM or GRU, but the dimensions are wrong resulting in:

Exception: Input 0 is incompatible with layer lstm_1: expected ndim=3, found ndim=2

I know the input shape for the LSTM wants a different ordering (i.e. to include the time dimension). 

I am not clear on how to reshape the data to pipe into the LSTM successfully from my Dense or Flatten layer.

Thanks in advance for any help,

CD

[p.nec...@gmail.com]

Doesn't reshape(1,dim) work?
http://keras.io/layers/core/#reshape

[CD]

Thanks

That did it

.. "duh" comes to mind...