python.closure

cond_closed_over_variable

Note

Tags: torch.cond, python.closure

Support Level: SUPPORTED

Original source code:

# mypy: allow-untyped-defs
import torch

from functorch.experimental.control_flow import cond

class CondClosedOverVariable(torch.nn.Module):
    """
    torch.cond() supports branches closed over arbitrary variables.
    """

    def forward(self, pred, x):
        def true_fn(val):
            return x * 2

        def false_fn(val):
            return x - 2

        return cond(pred, true_fn, false_fn, [x + 1])

example_args = (torch.tensor(True), torch.randn(3, 2))
tags = {"torch.cond", "python.closure"}
model = CondClosedOverVariable()


torch.export.export(model, example_args)

Result:

ExportedProgram:
    class GraphModule(torch.nn.Module):
        def forward(self, pred: "b8[]", x: "f32[3, 2]"):
                 true_graph_0 = self.true_graph_0
            false_graph_0 = self.false_graph_0
            cond = torch.ops.higher_order.cond(pred, true_graph_0, false_graph_0, [x]);  pred = true_graph_0 = false_graph_0 = x = None
            getitem: "f32[3, 2]" = cond[0];  cond = None
            return (getitem,)

        class true_graph_0(torch.nn.Module):
            def forward(self, x: "f32[3, 2]"):
                         mul: "f32[3, 2]" = torch.ops.aten.mul.Tensor(x, 2);  x = None
                return (mul,)

        class false_graph_0(torch.nn.Module):
            def forward(self, x: "f32[3, 2]"):
                         sub: "f32[3, 2]" = torch.ops.aten.sub.Tensor(x, 2);  x = None
                return (sub,)

Graph signature: ExportGraphSignature(input_specs=[InputSpec(kind=<InputKind.USER_INPUT: 1>, arg=TensorArgument(name='pred'), target=None, persistent=None), InputSpec(kind=<InputKind.USER_INPUT: 1>, arg=TensorArgument(name='x'), target=None, persistent=None)], output_specs=[OutputSpec(kind=<OutputKind.USER_OUTPUT: 1>, arg=TensorArgument(name='getitem'), target=None)])
Range constraints: {}

nested_function

Note

Tags: python.closure

Support Level: SUPPORTED

Original source code:

# mypy: allow-untyped-defs
import torch

class NestedFunction(torch.nn.Module):
    """
    Nested functions are traced through. Side effects on global captures
    are not supported though.
    """

    def forward(self, a, b):
        x = a + b
        z = a - b

        def closure(y):
            nonlocal x
            x += 1
            return x * y + z

        return closure(x)

example_args = (torch.randn(3, 2), torch.randn(2))
tags = {"python.closure"}
model = NestedFunction()


torch.export.export(model, example_args)

Result:

ExportedProgram:
    class GraphModule(torch.nn.Module):
        def forward(self, a: "f32[3, 2]", b: "f32[2]"):
                 add: "f32[3, 2]" = torch.ops.aten.add.Tensor(a, b)

                 sub: "f32[3, 2]" = torch.ops.aten.sub.Tensor(a, b);  a = b = None

                 add_1: "f32[3, 2]" = torch.ops.aten.add.Tensor(add, 1);  add = None

                 mul: "f32[3, 2]" = torch.ops.aten.mul.Tensor(add_1, add_1);  add_1 = None
            add_2: "f32[3, 2]" = torch.ops.aten.add.Tensor(mul, sub);  mul = sub = None
            return (add_2,)

Graph signature: ExportGraphSignature(input_specs=[InputSpec(kind=<InputKind.USER_INPUT: 1>, arg=TensorArgument(name='a'), target=None, persistent=None), InputSpec(kind=<InputKind.USER_INPUT: 1>, arg=TensorArgument(name='b'), target=None, persistent=None)], output_specs=[OutputSpec(kind=<OutputKind.USER_OUTPUT: 1>, arg=TensorArgument(name='add_2'), target=None)])
Range constraints: {}