for example, I have a tensorflow model like this:
node { name: “input0” op: “Placeholder” } node { name: “input1” op: “Placeholder” } node { name: “input2” op: “Placeholder” } … … node { name: “out0” op: “Add” } node { name: “out1” op: “Add” } node { name: “out2” op: “Add” }
this tf model has 3 inputs ,and 3 outputs,
then after convert to tvm, when I deploy it with c++,
Is it guaranteed that ,when I call set_input(0, data), it’s the same with set_input(‘input0’, data)？