# model model = Model() i1 = Input("op1", "TENSOR_FLOAT32", "{1, 2, 2, 1}") i2 = Input("op2", "TENSOR_FLOAT32", "{1, 2, 2, 1}") act = Int32Scalar("act", 0) # an int32_t scalar fuse_activation i3 = Output("op3", "TENSOR_FLOAT32", "{1, 2, 2, 1}") model = model.Operation("DIV", i1, i2, act).To(i3) # Example 1. Input in operand 0, input0 = {i1: # input 0 [2.0, -4.0, 8.0, -16.0], i2: # input 1 [2.0, -2.0, -4.0, 4.0]} output0 = {i3: # output 0 [1.0, 2.0, -2.0, -4.0]} # Instantiate an example Example((input0, output0)) # Test DIV by zero. # It is undefined behavior. The output is ignored and we only require the drivers to not crash. input0 = Input("input0", "TENSOR_FLOAT32", "{1}") input1 = Input("input1", "TENSOR_FLOAT32", "{1}") output = IgnoredOutput("output", "TENSOR_FLOAT32", "{1}") model = Model("by_zero").Operation("DIV", input0, input1, 0).To(output) Example({ input0: [1], input1: [0], output: [0], }).AddVariations("relaxed")