torch.export

Warning

This feature is a prototype under active development and there WILL BE BREAKING CHANGES in the future.

Overview

torch.export.export() takes an arbitrary Python callable (a torch.nn.Module, a function or a method) and produces a traced graph representing only the Tensor computation of the function in an Ahead-of-Time (AOT) fashion, which can subsequently be executed with different outputs or serialized.

import torch
from torch.export import export

def f(x: torch.Tensor, y: torch.Tensor) -> torch.Tensor:
    a = torch.sin(x)
    b = torch.cos(y)
    return a + b

example_args = (torch.randn(10, 10), torch.randn(10, 10))

exported_program: torch.export.ExportedProgram = export(
    f, args=example_args
)
print(exported_program)

ExportedProgram:
    class GraphModule(torch.nn.Module):
        def forward(self, arg0_1: f32[10, 10], arg1_1: f32[10, 10]):
            # code: a = torch.sin(x)
            sin: f32[10, 10] = torch.ops.aten.sin.default(arg0_1);

            # code: b = torch.cos(y)
            cos: f32[10, 10] = torch.ops.aten.cos.default(arg1_1);

            # code: return a + b
            add: f32[10, 10] = torch.ops.aten.add.Tensor(sin, cos);
            return (add,)

    Graph signature: ExportGraphSignature(
        parameters=[],
        buffers=[],
        user_inputs=['arg0_1', 'arg1_1'],
        user_outputs=['add'],
        inputs_to_parameters={},
        inputs_to_buffers={},
        buffers_to_mutate={},
        backward_signature=None,
        assertion_dep_token=None,
    )
    Range constraints: {}
    Equality constraints: []

torch.export produces a clean intermediate representation (IR) with the following invariants. More specifications about the IR can be found here (coming soon!).

• Soundness: It is guaranteed to be a sound representation of the original program, and maintains the same calling conventions of the original program.

• Normalized: There are no Python semantics within the graph. Submodules from the original programs are inlined to form one fully flattened computational graph.

• Defined Operator Set: The graph produced contains only a small defined Core ATen IR opset and registered custom operators.

• Graph properties: The graph is purely functional, meaning it does not contain operations with side effects such as mutations or aliasing. It does not mutate any intermediate values, parameters, or buffers.

• Metadata: The graph contains metadata captured during tracing, such as a stacktrace from user’s code.

Under the hood, torch.export leverages the following latest technologies:

• TorchDynamo (torch._dynamo) is an internal API that uses a CPython feature called the Frame Evaluation API to safely trace PyTorch graphs. This provides a massively improved graph capturing experience, with much fewer rewrites needed in order to fully trace the PyTorch code.

• AOT Autograd provides a functionalized PyTorch graph and ensures the graph is decomposed/lowered to the small defined Core ATen operator set.

• Torch FX (torch.fx) is the underlying representation of the graph, allowing flexible Python-based transformations.

Existing frameworks

torch.compile() also utilizes the same PT2 stack as torch.export, but is slightly different:

• JIT vs. AOT: torch.compile() is a JIT compiler whereas which is not intended to be used to produce compiled artifacts outside of deployment.

• Partial vs. Full Graph Capture: When torch.compile() runs into an untraceable part of a model, it will “graph break” and fall back to running the program in the eager Python runtime. In comparison, torch.export aims to get a full graph representation of a PyTorch model, so it will error out when something untraceable is reached. Since torch.export produces a full graph disjoint from any Python features or runtime, this graph can then be saved, loaded, and run in different environments and languages.

• Usability tradeoff: Since torch.compile() is able to fallback to the Python runtime whenever it reaches something untraceable, it is a lot more flexible. torch.export will instead require users to provide more information or rewrite their code to make it traceable.

Compared to torch.fx.symbolic_trace(), torch.export traces using TorchDynamo which operates at the Python bytecode level, giving it the ability to trace arbitrary Python constructs not limited by what Python operator overloading supports. Additionally, torch.export keeps fine-grained track of tensor metadata, so that conditionals on things like tensor shapes do not fail tracing. In general, torch.export is expected to work on more user programs, and produce lower-level graphs (at the torch.ops.aten operator level). Note that users can still use torch.fx.symbolic_trace() as a preprocessing step before torch.export.

Compared to torch.jit.script(), torch.export does not capture Python control flow or data structures, but it supports more Python language features than TorchScript (as it is easier to have comprehensive coverage over Python bytecodes). The resulting graphs are simpler and only have straight line control flow (except for explicit control flow operators).

Compared to torch.jit.trace(), torch.export is sound: it is able to trace code that performs integer computation on sizes and records all of the side-conditions necessary to show that a particular trace is valid for other inputs.

Exporting a PyTorch Model

An Example

The main entrypoint is through torch.export.export(), which takes a callable (torch.nn.Module, function, or method) and sample inputs, and captures the computation graph into an torch.export.ExportedProgram. An example:

import torch
from torch.export import export

# Simple module for demonstration
class M(torch.nn.Module):
    def __init__(self) -> None:
        super().__init__()
        self.conv = torch.nn.Conv2d(
            in_channels=3, out_channels=16, kernel_size=3, padding=1
        )
        self.relu = torch.nn.ReLU()
        self.maxpool = torch.nn.MaxPool2d(kernel_size=3)

    def forward(self, x: torch.Tensor, *, constant=None) -> torch.Tensor:
        a = self.conv(x)
        a.add_(constant)
        return self.maxpool(self.relu(a))

example_args = (torch.randn(1, 3, 256, 256),)
example_kwargs = {"constant": torch.ones(1, 16, 256, 256)}

exported_program: torch.export.ExportedProgram = export(
    M(), args=example_args, kwargs=example_kwargs
)
print(exported_program)

ExportedProgram:
    class GraphModule(torch.nn.Module):
        def forward(self, arg0_1: f32[16, 3, 3, 3], arg1_1: f32[16], arg2_1: f32[1, 3, 256, 256], arg3_1: f32[1, 16, 256, 256]):

            # code: a = self.conv(x)
            convolution: f32[1, 16, 256, 256] = torch.ops.aten.convolution.default(
                arg2_1, arg0_1, arg1_1, [1, 1], [1, 1], [1, 1], False, [0, 0], 1
            );

            # code: a.add_(constant)
            add: f32[1, 16, 256, 256] = torch.ops.aten.add.Tensor(convolution, arg3_1);

            # code: return self.maxpool(self.relu(a))
            relu: f32[1, 16, 256, 256] = torch.ops.aten.relu.default(add);
            max_pool2d_with_indices = torch.ops.aten.max_pool2d_with_indices.default(
                relu, [3, 3], [3, 3]
            );
            getitem: f32[1, 16, 85, 85] = max_pool2d_with_indices[0];
            return (getitem,)

    Graph signature: ExportGraphSignature(
        parameters=['L__self___conv.weight', 'L__self___conv.bias'],
        buffers=[],
        user_inputs=['arg2_1', 'arg3_1'],
        user_outputs=['getitem'],
        inputs_to_parameters={
            'arg0_1': 'L__self___conv.weight',
            'arg1_1': 'L__self___conv.bias',
        },
        inputs_to_buffers={},
        buffers_to_mutate={},
        backward_signature=None,
        assertion_dep_token=None,
    )
    Range constraints: {}
    Equality constraints: []

Inspecting the ExportedProgram, we can note the following:

• The torch.fx.Graph contains the computation graph of the original program, along with records of the original code for easy debugging.

• The graph contains only torch.ops.aten operators found in the Core ATen IR opset and custom operators, and is fully functional, without any inplace operators such as torch.add_.

• The parameters (weight and bias to conv) are lifted as inputs to the graph, resulting in no get_attr nodes in the graph, which previously existed in the result of torch.fx.symbolic_trace().

• The torch.export.ExportGraphSignature models the input and output signature, along with specifying which inputs are parameters.

• The resulting shape and dtype of tensors produced by each node in the graph is noted. For example, the convolution node will result in a tensor of dtype torch.float32 and shape (1, 16, 256, 256).

Expressing Dynamism

By default torch.export will trace the program assuming all input shapes are static, and specializing the exported program to those dimensions. However, some dimensions, such as a batch dimension, can be dynamic and vary from run to run. Such dimensions must be marked dynamic using the torch.export.dynamic_dim() API, and passed into torch.export.export() through the constraints argument. An example:

import torch
from torch.export import export, dynamic_dim

class M(torch.nn.Module):
    def __init__(self):
        super().__init__()

        self.branch1 = torch.nn.Sequential(
            torch.nn.Linear(64, 32), torch.nn.ReLU()
        )
        self.branch2 = torch.nn.Sequential(
            torch.nn.Linear(128, 64), torch.nn.ReLU()
        )
        self.buffer = torch.ones(32)

    def forward(self, x1, x2):
        out1 = self.branch1(x1)
        out2 = self.branch2(x2)
        return (out1 + self.buffer, out2)

example_args = (torch.randn(32, 64), torch.randn(32, 128))
constraints = [
    # First dimension of each input is a dynamic batch size
    dynamic_dim(example_args[0], 0),
    dynamic_dim(example_args[1], 0),
    # The dynamic batch size between the inputs are equal
    dynamic_dim(example_args[0], 0) == dynamic_dim(example_args[1], 0),
]

exported_program: torch.export.ExportedProgram = export(
  M(), args=example_args, constraints=constraints
)
print(exported_program)

ExportedProgram:
    class GraphModule(torch.nn.Module):
        def forward(self, arg0_1: f32[32, 64], arg1_1: f32[32], arg2_1: f32[64, 128], arg3_1: f32[64], arg4_1: f32[32], arg5_1: f32[s0, 64], arg6_1: f32[s0, 128]):

            # code: out1 = self.branch1(x1)
            permute: f32[64, 32] = torch.ops.aten.permute.default(arg0_1, [1, 0]);
            addmm: f32[s0, 32] = torch.ops.aten.addmm.default(arg1_1, arg5_1, permute);
            relu: f32[s0, 32] = torch.ops.aten.relu.default(addmm);

            # code: out2 = self.branch2(x2)
            permute_1: f32[128, 64] = torch.ops.aten.permute.default(arg2_1, [1, 0]);
            addmm_1: f32[s0, 64] = torch.ops.aten.addmm.default(arg3_1, arg6_1, permute_1);
            relu_1: f32[s0, 64] = torch.ops.aten.relu.default(addmm_1);  addmm_1 = None

            # code: return (out1 + self.buffer, out2)
            add: f32[s0, 32] = torch.ops.aten.add.Tensor(relu, arg4_1);
            return (add, relu_1)

    Graph signature: ExportGraphSignature(
        parameters=[
            'branch1.0.weight',
            'branch1.0.bias',
            'branch2.0.weight',
            'branch2.0.bias',
        ],
        buffers=['L__self___buffer'],
        user_inputs=['arg5_1', 'arg6_1'],
        user_outputs=['add', 'relu_1'],
        inputs_to_parameters={
            'arg0_1': 'branch1.0.weight',
            'arg1_1': 'branch1.0.bias',
            'arg2_1': 'branch2.0.weight',
            'arg3_1': 'branch2.0.bias',
        },
        inputs_to_buffers={'arg4_1': 'L__self___buffer'},
        buffers_to_mutate={},
        backward_signature=None,
        assertion_dep_token=None,
    )
    Range constraints: {s0: RangeConstraint(min_val=2, max_val=9223372036854775806)}
    Equality constraints: [(InputDim(input_name='arg5_1', dim=0), InputDim(input_name='arg6_1', dim=0))]

Some additional things to note:

• Through the torch.export.dynamic_dim() API, we specified the first dimension of each input to be dynamic. Looking at the inputs arg5_1 and arg6_1, they have a symbolic shape of (s0, 64) and (s0, 128), instead of the (32, 64) and (32, 128) shaped tensors that we passed in as example inputs. s0 is a symbol representing that this dimension can be a range of values.

• exported_program.range_constraints describes the ranges of each symbol appearing in the graph. In this case, we see that s0 has the range [2, inf]. For technical reasons that are difficult to explain here, they are assumed to be not 0 or 1. This is not a bug, and does not necessarily mean that the exported program will not work for dimensions 0 or 1. See The 0/1 Specialization Problem for an in-depth discussion of this topic.

• exported_program.equality_constraints describes which dimensions are required to be equal. Since we specified in the constraints that the first dimension of each argument is equivalent, (dynamic_dim(example_args[0], 0) == dynamic_dim(example_args[1], 0)), we see in the equality constraints the tuple specifying that arg5_1 dimension 0 and arg6_1 dimension 0 are equal.

Serialization

To save the ExportedProgram, users can use the torch.export.save() and torch.export.load() APIs. A convention is to save the ExportedProgram using a .pt2 file extension.

An example:

import torch
import io

class MyModule(torch.nn.Module):
    def forward(self, x):
        return x + 10

exported_program = torch.export.export(MyModule(), torch.randn(5))

torch.export.save(exported_program, 'exported_program.pt2')
saved_exported_program = torch.export.load('exported_program.pt2')

Specialization

Input shapes

As mentioned before, by default, torch.export will trace the program specializing on the input tensors’ shapes, unless a dimension is specified as dynamic via the torch.export.dynamic_dim() API. This means that if there exists shape-dependent control flow, torch.export will specialize on the branch that is being taken with the given sample inputs. For example:

import torch
from torch.export import export

def fn(x):
    if x.shape[0] > 5:
        return x + 1
    else:
        return x - 1

example_inputs = (torch.rand(10, 2),)
exported_program = export(fn, example_inputs)
print(exported_program)

ExportedProgram:
    class GraphModule(torch.nn.Module):
        def forward(self, arg0_1: f32[10, 2]):
            add: f32[10, 2] = torch.ops.aten.add.Tensor(arg0_1, 1);
            return (add,)

The conditional of (x.shape[0] > 5) does not appear in the ExportedProgram because the example inputs have the static shape of (10, 2). Since torch.export specializes on the inputs’ static shapes, the else branch (x - 1) will never be reached. To preserve the dynamic branching behavior based on the shape of a tensor in the traced graph, torch.export.dynamic_dim() will need to be used to specify the dimension of the input tensor (x.shape[0]) to be dynamic, and the source code will need to be rewritten.

Non-tensor inputs

torch.export also specializes the traced graph based on the values of inputs that are not torch.Tensor, such as int, float, bool, and str. However, we will likely change this in the near future to not specialize on inputs of primitive types.

For example:

import torch
from torch.export import export

def fn(x: torch.Tensor, const: int, times: int):
    for i in range(times):
        x = x + const
    return x

example_inputs = (torch.rand(2, 2), 1, 3)
exported_program = export(fn, example_inputs)
print(exported_program)

ExportedProgram:
    class GraphModule(torch.nn.Module):
        def forward(self, arg0_1: f32[2, 2], arg1_1, arg2_1):
            add: f32[2, 2] = torch.ops.aten.add.Tensor(arg0_1, 1);
            add_1: f32[2, 2] = torch.ops.aten.add.Tensor(add, 1);
            add_2: f32[2, 2] = torch.ops.aten.add.Tensor(add_1, 1);
            return (add_2,)

Because integers are specialized, the torch.ops.aten.add.Tensor operations are all computed with the inlined constant 1, rather than arg1_1. Additionally, the times iterator used in the for loop is also “inlined” in the graph through the 3 repeated torch.ops.aten.add.Tensor calls, and the input arg2_1 is never used.

Limitations of torch.export

Graph Breaks

As torch.export is a one-shot process for capturing a computation graph from a PyTorch program, it might ultimately run into untraceable parts of programs as it is nearly impossible to support tracing all PyTorch and Python features. In the case of torch.compile, an unsupported operation will cause a “graph break” and the unsupported operation will be run with default Python evaluation. In contrast, torch.export will require users to provide additional information or rewrite parts of their code to make it traceable. As the tracing is based on TorchDynamo, which evaluates at the Python bytecode level, there will be significantly fewer rewrites required compared to previous tracing frameworks.

When a graph break is encountered, ExportDB is a great resource for learning about the kinds of programs that are supported and unsupported, along with ways to rewrite programs to make them traceable.

Data/Shape-Dependent Control Flow

Graph breaks can also be encountered on data-dependent control flow (if x.shape[0] > 2) when shapes are not being specialized, as a tracing compiler cannot possibly deal with without generating code for a combinatorially exploding number of paths. In such cases, users will need to rewrite their code using special control flow operators (coming soon!).

Data-Dependent Accesses

Data dependent behavior such as using the value inside of a tensor to construct another tensor, or using the value of a tensor to slice into another tensor, is also something the tracer cannot fully determine. Users will need to rewrite their code using the inline constraint APIs torch.export.constrain_as_size() and torch.export.constrain_as_value().

Missing Meta Kernels for Operators

When tracing, a META implementation (or “meta kernel”) is required for all operators. This is used to reason about the input/output shapes for this operator.

Note that the official API for registering custom meta kernels for custom ops is currently undergoing development. While the final API is being refined, you can refer to the documentation here.

In the unfortunate case where your model uses an ATen operator that is does not have a meta kernel implementation yet, please file an issue.

Read More

Additional Links for Export Users

• Writing Graph Transformations on ATen IR
• IRs
• ExportDB

Deep Dive for PyTorch Developers

• TorchDynamo Deep Dive
• Dynamic shapes
• Fake tensor

API Reference

torch.export.export(f, args, kwargs=None, *, constraints=None)

export() takes an arbitrary Python callable (an nn.Module, a function or a method) and produces a traced graph representing only the Tensor computation of the function in an Ahead-of-Time (AOT) fashion, which can subsequently be executed with different outputs or serialized. The traced graph (1) produces a normalized operator set consisting only of functional Core ATen Operator Set and user specified custom operators, (2) has eliminated all Python control flow and data structures (except for certain conditions), and (3) has the set of shape constraints needed to show that this normalization and control flow elimination is sound for a future input.

Soundness Guarantee

While tracing, export() takes note of shape-related assumptions made by the user program and the underlying PyTorch operator kernels. The output ExportedProgram is considered valid only when these assumptions hold true.

There are 2 types of assumptions made during tracing

• Shapes (not values) of input tensors.

• Ranges (lower and upper bound) of values extracted from intermediate tensors via .item() or direct indexing.

All assumptions must be validated at graph capture time for export() to succeed. Specifically:

• Assumptions on static shapes of input tensors are automatically validated without additional effort.

• Assumptions on dynamic shape of input tensors require explicit Input Constraint constructed with dynamic_dim() APIs

• Assumptions on range of intermediate values require explicit Inline Constraint, constructed use constrain_as_size() and constraint_as_value() APIs.

If any assumption can not be validated, a fatal error will be raised. When that happens, the error message will include suggested code needed to construct necessary constraints to validate the assumptions, for example export() would suggest following code for input constraints:

def specify_constraints(x):
    return [
        # x:
        dynamic_dim(x, 0) <= 5,
    ]

This example means the program requires the dim 0 of input x to be less than or equal to 5 to be valid. You can inspect the constraints needed and then copy this exact function into your code to generated needed constraints to be passed into constraints argument.

Parameters

• f (Callable) – The callable to trace.

• args (Tuple[Any, ...]) – Example positional inputs.

• kwargs (Optional[Dict[str, Any]]) – Optional example keyword inputs.

• constraints (Optional[List[Constraint]]) – An optional list of constraints on the dynamic arguments that specify their possible range of shapes. By default, shapes of input torch.Tensors are assumed to be static. If an input torch.Tensor is expected to have dynamic shapes, please use dynamic_dim() to define Constraint objects that specify the dynamics and the possible range of shapes. See dynamic_dim() docstring for examples on how to use it.

Returns

An ExportedProgram containing the traced callable.

Return type

ExportedProgram

Acceptable input/output types

Acceptable types of inputs (for args and kwargs) and outputs include:

• Primitive types, i.e. torch.Tensor, int, float, bool and str.

• (Nested) Data structures comprising of dict, list, tuple, namedtuple and OrderedDict containing all above types.

torch.export.dynamic_dim(t, index)

dynamic_dim() constructs a Constraint object that describes the dynamism of a dimension index of tensor t. Constraint objects should be passed to constraints argument of export().

Parameters

• t (torch.Tensor) – Example input tensor that have dynamic dimension size(s)

• index (int) – Index of dynamic dimension

Returns

A Constraint object that describes shape dynamism. It can be passed to export() so that export() does not assume static size of specified tensor, i.e. keeping it dynamic as a symbolic size rather than specializing according to size of example tracing input.

Specifically dynamic_dim() can be used to express following types of dynamism.

• Size of a dimension is dynamic and unbounded:

  t0 = torch.rand(2, 3)
  t1 = torch.rand(3, 4)
  
  # First dimension of t0 can be dynamic size rather than always being static size 2
  constraints = [dynamic_dim(t0, 0)]
  ep = export(fn, (t0, t1), constraints=constraints)
  

• Size of a dimension is dynamic with a lower bound:

  t0 = torch.rand(10, 3)
  t1 = torch.rand(3, 4)
  
  # First dimension of t0 can be dynamic size with a lower bound of 5 (inclusive)
  # Second dimension of t1 can be dynamic size with a lower bound of 2 (exclusive)
  constraints = [
      dynamic_dim(t0, 0) >= 5,
      dynamic_dim(t1, 1) > 2,
  ]
  ep = export(fn, (t0, t1), constraints=constraints)
  

• Size of a dimension is dynamic with an upper bound:

  t0 = torch.rand(10, 3)
  t1 = torch.rand(3, 4)
  
  # First dimension of t0 can be dynamic size with a upper bound of 16 (inclusive)
  # Second dimension of t1 can be dynamic size with a upper bound of 8 (exclusive)
  constraints = [
      dynamic_dim(t0, 0) <= 16,
      dynamic_dim(t1, 1) < 8,
  ]
  ep = export(fn, (t0, t1), constraints=constraints)
  

• Size of a dimension is dynamic and it is always equal to size of another dynamic dimension:

  t0 = torch.rand(10, 3)
  t1 = torch.rand(3, 4)
  
  # Sizes of second dimension of t0 and first dimension are always equal
  constraints = [
      dynamic_dim(t0, 1) == dynamic_dim(t1, 0),
  ]
  ep = export(fn, (t0, t1), constraints=constraints)
  

• Mix and match all types above as long as they do not express conflicting requirements

torch.export.constrain_as_size(symbol, min=None, max=None)

Hint export() about the constraint of an intermediate scalar value that represents shape of a tensor so that subsequent tensor constructors can be traced correctly because many operators need to make assumption about range of sizes.

Parameters

• symbol – Intermediate scalar value (int-only now) to apply range constraint on.

• min (Optional[int]) – Minimum possible value of given symbol (inclusive)

• max (Optional[int]) – Maximum possible value of given symbol (inclusive)

Returns

None

For example, following program can not be traced soundly wihout using constrain_as_size() to give export() a hint about shape ranges:

def fn(x):
    d = x.max().item()
    return torch.ones(v)

export() would give following error:

torch._dynamo.exc.Unsupported: guard on data-dependent symbolic int/float

Assuming the actual range of d can be between [3, 10], you can add a call to constrain_as_size() in the source code like this:

def fn(x):
    d = x.max().item()
    torch.export.constrain_as_size(d, min=3, max=10)
    return torch.ones(d)

With the additional hint, export() would be able to trace the program correctly by taking the else branch, resulting in following graph:

graph():
    %arg0_1 := placeholder[target=arg0_1]

    # d = x.max().item()
    %max_1 := call_function[target=torch.ops.aten.max.default](args = (%arg0_1,))
    %_local_scalar_dense := call_function[target=torch.ops.aten._local_scalar_dense.default](args = (%max_1,))

    # Asserting 3 <= d <= 10
    %ge := call_function[target=operator.ge](args = (%_local_scalar_dense, 3))
    %scalar_tensor := call_function[target=torch.ops.aten.scalar_tensor.default](args = (%ge,))
    %_assert_async := call_function[target=torch.ops.aten._assert_async.msg](
        args = (%scalar_tensor, _local_scalar_dense is outside of inline constraint [3, 10].))
    %le := call_function[target=operator.le](args = (%_local_scalar_dense, 10))
    %scalar_tensor_1 := call_function[target=torch.ops.aten.scalar_tensor.default](args = (%le,))
    %_assert_async_1 := call_function[target=torch.ops.aten._assert_async.msg](
        args = (%scalar_tensor_1, _local_scalar_dense is outside of inline constraint [3, 10].))
    %sym_constrain_range_for_size := call_function[target=torch.ops.aten.sym_constrain_range_for_size.default](
        args = (%_local_scalar_dense,), kwargs = {min: 3, max: 10})

    # Constructing new tensor with d
    %full := call_function[target=torch.ops.aten.full.default](
        args = ([%_local_scalar_dense], 1),
        kwargs = {dtype: torch.float32, layout: torch.strided, device: cpu, pin_memory: False})

    ......

Warning

if your size is intended to be dynamic, do NOT test if sizes are equal to 0 or 1, these will SILENTLY report false and be bypassed

torch.export.constrain_as_value(symbol, min=None, max=None)

Hint export() about the constraint of an intermediate scalar value so that subsequent branching behaviors that check on the range of aforementioned scalar value can be soundly traced.

Warning

(Note that if the intermediate scalar value will be used like a size, including being passed as size arg to a tensor factory or view, call constrain_as_size() instead.)

Parameters

• symbol – Intermediate scalar value (int-only now) to apply range constraint on.

• min (Optional[int]) – Minimum possible value of given symbol (inclusive)

• max (Optional[int]) – Maximum possible value of given symbol (inclusive)

Returns

None

For example, following program can not be traced soundly:

def fn(x):
    v = x.max().item()
    if v > 1024:
        return x
    else:
        return x * 2

v is a data-dependent value, which is assumed to have a range of (-inf, inf). export() a hint about which branch to take would not be able to determine if the traced branching decision is correct or not. Thus export() would give following error:

torch._dynamo.exc.UserError: Consider annotating your code using
torch.export.constrain_as_size() or torch.export().constrain_as_value() APIs.
It appears that you're trying to get a value out of symbolic int/float whose value
is data-dependent (and thus we do not know the true value.)  The expression we were
trying to evaluate is f0 > 1024 (unhinted: f0 > 1024).

Assuming the actual range of v can be between [10, 200], you can add a call to constrain_as_value() in the source code like this:

def fn(x):
    v = x.max().item()

    # Give export() a hint
    torch.export.constrain_as_value(v, min=10, max=200)

    if v > 1024:
        return x
    else:
        return x * 2

With the additional hint, export() would be able to trace the program correctly by taking the else branch, resulting in following graph:

graph():
    %arg0_1 := placeholder[target=arg0_1]

    # v = x.max().item()
    %max_1 := call_function[target=torch.ops.aten.max.default](args = (%arg0_1,))
    %_local_scalar_dense := call_function[target=torch.ops.aten._local_scalar_dense.default](args = (%max_1,))

    # Asserting 10 <= v <= 200
    %ge := call_function[target=operator.ge](args = (%_local_scalar_dense, 10))
    %scalar_tensor := call_function[target=torch.ops.aten.scalar_tensor.default](args = (%ge,))
    %_assert_async := call_function[target=torch.ops.aten._assert_async.msg](
        args = (%scalar_tensor, _local_scalar_dense is outside of inline constraint [10, 200].))
    %le := call_function[target=operator.le](args = (%_local_scalar_dense, 200))
    %scalar_tensor_1 := call_function[target=torch.ops.aten.scalar_tensor.default](args = (%le,))
    %_assert_async_1 := call_function[target=torch.ops.aten._assert_async.msg](
        args = (%scalar_tensor_1, _local_scalar_dense is outside of inline constraint [10, 200].))
    %sym_constrain_range := call_function[target=torch.ops.aten.sym_constrain_range.default](
        args = (%_local_scalar_dense,), kwargs = {min: 10, max: 200})

    # Always taking `else` branch to multiply elements `x` by 2 due to hints above
    %mul := call_function[target=torch.ops.aten.mul.Tensor](args = (%arg0_1, 2), kwargs = {})
    return (mul,)

torch.export.save(ep, f, *, extra_files=None, opset_version=None)

Warning

Under active development, saved files may not be usable in newer versions of PyTorch.

Saves an ExportedProgram to a file-like object. It can then be loaded using the Python API torch.export.load.

Parameters

• ep (ExportedProgram) – The exported program to save.

• f (Union[str, pathlib.Path, io.BytesIO) – A file-like object (has to implement write and flush) or a string containing a file name.

• extra_files (Optional[Dict[str, Any]]) – Map from filename to contents which will be stored as part of f.

• opset_version (Optional[Dict[str, int]]) – A map of opset names to the version of this opset

Example:

import torch
import io

class MyModule(torch.nn.Module):
    def forward(self, x):
        return x + 10

ep = torch.export.export(MyModule(), torch.randn(5))

# Save to file
torch.export.save(ep, 'exported_program.pt2')

# Save to io.BytesIO buffer
buffer = io.BytesIO()
torch.export.save(ep, buffer)

# Save with extra files
extra_files = {'foo.txt': b'bar'}
torch.export.save(ep, 'exported_program.pt2', extra_files=extra_files)

torch.export.load(f, *, extra_files=None, expected_opset_version=None)

Warning

Under active development, saved files may not be usable in newer versions of PyTorch.

Loads an ExportedProgram previously saved with torch.export.save.

Parameters

• ep (ExportedProgram) – The exported program to save.

• f (Union[str, pathlib.Path, io.BytesIO) – A file-like object (has to implement write and flush) or a string containing a file name.

• extra_files (Optional[Dict[str, Any]]) – The extra filenames given in this map would be loaded and their content would be stored in the provided map.

• expected_opset_version (Optional[Dict[str, int]]) – A map of opset names to expected opset versions

Returns

An ExportedProgram object

Return type

ExportedProgram

Example:

import torch
import io

# Load ExportedProgram from file
ep = torch.export.load('exported_program.pt2')

# Load ExportedProgram from io.BytesIO object
with open('exported_program.pt2', 'rb') as f:
    buffer = io.BytesIO(f.read())
buffer.seek(0)
ep = torch.export.load(buffer)

# Load with extra files.
extra_files = {'foo.txt': ''}  # values will be replaced with data
ep = torch.export.load('exported_program.pt2', extra_files=extra_files)
print(extra_files['foo.txt'])

class torch.export.Constraint(*args, **kwargs)

Warning

Do not construct Constraint directly, use dynamic_dim() instead.

This represents constraints on input tensor dimensions, e.g., requiring them to be fully polymorphic or within some range.

class torch.export.ExportedProgram(root, graph, graph_signature, call_spec, state_dict, range_constraints, equality_constraints, module_call_graph, example_inputs=None)

Package of a program from export(). It contains an torch.fx.Graph that represents Tensor computation, a state_dict containing tensor values of all lifted parameters and buffers, and various metadata.

You can call an ExportedProgram like the original callable traced by export() with the same calling convention.

To perform transformations on the graph, use .module property to access an torch.fx.GraphModule. You can then use FX transformation to rewrite the graph. Afterwards, you can simply use export() again to construct a correct ExportedProgram.

module()

Returns a self contained GraphModule with all the parameters/buffers inlined.

Return type

Module

class torch.export.ExportBackwardSignature(gradients_to_parameters: Dict[str, str], gradients_to_user_inputs: Dict[str, str], loss_output: str)

class torch.export.ExportGraphSignature(parameters, buffers, user_inputs, user_outputs, inputs_to_parameters, inputs_to_buffers, buffers_to_mutate, backward_signature, assertion_dep_token=None)

ExportGraphSignature models the input/output signature of Export Graph, which is a fx.Graph with stronger invariants gurantees.

Export Graph is functional and does not access “states” like parameters or buffers within the graph via getattr nodes. Instead, export() gurantees that parameters and buffers are lifted out of the graph as inputs. Similarly, any mutations to buffers are not included in the graph either, instead the updated values of mutated buffers are modeled as additional outputs of Export Graph.

The ordering of all inputs and outputs are:

Inputs = [*parameters_buffers, *flattened_user_inputs]
Outputs = [*mutated_inputs, *flattened_user_outputs]

e.g. If following module is exported:

class CustomModule(nn.Module):
    def __init__(self):
        super(CustomModule, self).__init__()

        # Define a parameter
        self.my_parameter = nn.Parameter(torch.tensor(2.0))

        # Define two buffers
        self.register_buffer('my_buffer1', torch.tensor(3.0))
        self.register_buffer('my_buffer2', torch.tensor(4.0))

    def forward(self, x1, x2):
        # Use the parameter, buffers, and both inputs in the forward method
        output = (x1 + self.my_parameter) * self.my_buffer1 + x2 * self.my_buffer2

        # Mutate one of the buffers (e.g., increment it by 1)
        self.my_buffer2.add_(1.0) # In-place addition

        return output

Resulting Graph would be:

graph():
    %arg0_1 := placeholder[target=arg0_1]
    %arg1_1 := placeholder[target=arg1_1]
    %arg2_1 := placeholder[target=arg2_1]
    %arg3_1 := placeholder[target=arg3_1]
    %arg4_1 := placeholder[target=arg4_1]
    %add_tensor := call_function[target=torch.ops.aten.add.Tensor](args = (%arg3_1, %arg0_1), kwargs = {})
    %mul_tensor := call_function[target=torch.ops.aten.mul.Tensor](args = (%add_tensor, %arg1_1), kwargs = {})
    %mul_tensor_1 := call_function[target=torch.ops.aten.mul.Tensor](args = (%arg4_1, %arg2_1), kwargs = {})
    %add_tensor_1 := call_function[target=torch.ops.aten.add.Tensor](args = (%mul_tensor, %mul_tensor_1), kwargs = {})
    %add_tensor_2 := call_function[target=torch.ops.aten.add.Tensor](args = (%arg2_1, 1.0), kwargs = {})
    return (add_tensor_2, add_tensor_1)

Resulting ExportGraphSignature would be:

ExportGraphSignature(
    # Indicates that there is one parameter named `my_parameter`
    parameters=['L__self___my_parameter'],

    # Indicates that there are two buffers, `my_buffer1` and `my_buffer2`
    buffers=['L__self___my_buffer1', 'L__self___my_buffer2'],

    # Indicates that the nodes `arg3_1` and `arg4_1` in produced graph map to
    # original user inputs, ie. x1 and x2
    user_inputs=['arg3_1', 'arg4_1'],

    # Indicates that the node `add_tensor_1` maps to output of original program
    user_outputs=['add_tensor_1'],

    # Indicates that there is one parameter (self.my_parameter) captured,
    # its name is now mangled to be `L__self___my_parameter`, which is now
    # represented by node `arg0_1` in the graph.
    inputs_to_parameters={'arg0_1': 'L__self___my_parameter'},

    # Indicates that there are two buffers (self.my_buffer1, self.my_buffer2) captured,
    # their name are now mangled to be `L__self___my_my_buffer1` and `L__self___my_buffer2`.
    # They are now represented by nodes `arg1_1` and `arg2_1` in the graph.
    inputs_to_buffers={'arg1_1': 'L__self___my_buffer1', 'arg2_1': 'L__self___my_buffer2'},

    # Indicates that one buffer named `L__self___my_buffer2` is mutated during execution,
    # its new value is output from the graph represented by the node named `add_tensor_2`
    buffers_to_mutate={'add_tensor_2': 'L__self___my_buffer2'},

    # Backward graph not captured
    backward_signature=None,

    # Work in progress feature, please ignore now.
    assertion_dep_token=None
)

class torch.export.ArgumentKind(value)

An enumeration.

class torch.export.ArgumentSpec(kind: torch.export.ArgumentKind, value: Any)

class torch.export.ModuleCallSignature(inputs: List[torch.export.ArgumentSpec], outputs: List[torch.export.ArgumentSpec], in_spec: torch.utils._pytree.TreeSpec, out_spec: torch.utils._pytree.TreeSpec)

class torch.export.ModuleCallEntry(fqn: str, signature: Union[torch.export.ModuleCallSignature, NoneType] = None)