Closed AparaV closed 6 years ago
If you would like TensorFlow to automatically choose an existing and supported device to run the operations in case the specified one doesn't exist, you can set
allow_soft_placement
toTrue
in the configuration option when creating the session.
Taken from documentation
Implemented this by passing in a boolean value that chose between CPU and GPU devices. Code breaks during backprop. The other two graphs seem to work fine.