A DataflowGraph object is a high-level representation of the graph structure behind a DeepLearning model. By design, the construction of the model (the dataflow graph) is completely decoupled from any particular execution of the model, which is done with a Session object.

The default dataflow graph can be obtained by calling the GetDefaultGraph command.

$\mathrm{with}\left(\mathrm{DeepLearning}\right)\:$

$\mathrm{GetDefaultGraph}\left(\right)$

$\mathrm{Vector}\left(\left[\mathrm{DeepLearning\; DataflowGraph}\&comma;\mathrm{<tensorflow.python.framework.ops.Graph\; object\; at\; 0x7fc21c787898>}\right]\right)$

The DeepLearning[DataflowGraph] command was introduced in Maple 2018.

For more information on Maple 2018 changes, see Updates in Maple 2018.