f4001171bc0e65b736ceb55579fbacfcc64c4e85,dlpy/network.py,Network,__init__,#Network#Any#Any#Any#Any#Any#,67

Before Change

    name = "model" + str(number_of_instances)

    def __init__(self, conn, inputs=None, outputs=None, model_table=None, model_weights=None):
        if model_table is not None and any(i is not None for i in [inputs, outputs]):
            raise DLPyError("Either parameter model_table or inputs and outputs needs to be set.\n"
                            "The following cases are valid.\n"
                            "1. model_table = "your_model_table"; inputs = None; outputs = None.\n"
                            "2. model_table = None; inputs = input_layer(s); outputs = output_layer."
                            )
        self._init_model(conn, model_table, model_weights)
        if all(i is None for i in [inputs, outputs, model_table]):
            return
        if self.__class__.__name__ == "Model":
            if None in [inputs, outputs]:
                raise DLPyError("Parameter inputs and outputs are required.")
            self._map_graph_network(inputs, outputs)

    def _init_model(self, conn, model_table=None, model_weights=None):
        conn.loadactionset(actionSet="deeplearn", _messagelevel="error")

After Change

    name = None

    def __init__(self, conn, inputs=None, outputs=None, model_table=None, model_weights=None):
        if (inputs is None or outputs is None) and (inputs is not None or outputs is not None):
            raise ValueError("If one of inputs and outputs option is enabled, both should be specified")
        self._init_model(conn, model_table, model_weights)
        // works for Sequential() as well
        if self.__class__.__name__ == "Model":
            // 1). Model(s, model_table, model_weights)

Instances