API#
- class onnxruntime.training.ORTModule(module: Module, debug_options: Optional[DebugOptions] = None)[source]#
Bases:
Module
Extends user’s
torch.nn.Module
model to leverage ONNX Runtime super fast training engine.ORTModule specializes the user’s
torch.nn.Module
model, providingforward()
,backward()
along with all otherstorch.nn.Module
’s APIs.- Parameters:
module (torch.nn.Module) – User’s PyTorch module that ORTModule specializes
debug_options (
DebugOptions
, optional) – debugging options for ORTModule.
Initializes internal Module state, shared by both nn.Module and ScriptModule.
- forward(*inputs, **kwargs)[source]#
Delegate the
forward()
pass of PyTorch training to ONNX Runtime.The first call to forward performs setup and checking steps. During this call, ORTModule determines whether the module can be trained with ONNX Runtime. For this reason, the first forward call execution takes longer than subsequent calls. Execution is interupted if ONNX Runtime cannot process the model for training.
- Parameters:
inputs – positional, variable positional inputs to the PyTorch module’s forward method.
kwargs – keyword and variable keyword arguments to the PyTorch module’s forward method.
- Returns:
The output as expected from the forward method defined by the user’s PyTorch module. Output values supported include tensors, nested sequences of tensors and nested dictionaries of tensor values.
- add_module(name: str, module: Optional[Module]) None [source]#
Raises a ORTModuleTorchModelException exception since ORTModule does not support adding modules to it
- property module#
The original torch.nn.Module that this module wraps.
This property provides access to methods and properties on the original module.
- apply(fn: Callable[[Module], None]) T [source]#
Override
apply()
to delegate execution to ONNX Runtime
- state_dict(destination=None, prefix='', keep_vars=False)[source]#
Override
state_dict()
to delegate execution to ONNX Runtime
- load_state_dict(state_dict: OrderedDict[str, Tensor], strict: bool = True)[source]#
Override
load_state_dict()
to delegate execution to ONNX Runtime
- register_buffer(name: str, tensor: Optional[Tensor], persistent: bool = True) None [source]#
Override
register_buffer()
- register_parameter(name: str, param: Optional[Parameter]) None [source]#
Override
register_parameter()
- get_parameter(target: str) Parameter [source]#
Override
get_parameter()
- get_buffer(target: str) Tensor [source]#
Override
get_buffer()
- named_parameters(prefix: str = '', recurse: bool = True) Iterator[Tuple[str, Parameter]] [source]#
Override
named_parameters()
- named_buffers(prefix: str = '', recurse: bool = True) Iterator[Tuple[str, Tensor]] [source]#
Override
named_buffers()
- named_modules(*args, **kwargs)[source]#
Override
named_modules()
- bfloat16() T #
Casts all floating point parameters and buffers to
bfloat16
datatype.Note
This method modifies the module in-place.
- Returns:
self
- Return type:
- children() Iterator[Module] #
Returns an iterator over immediate children modules.
- Yields:
Module – a child module
- compile(*args, **kwargs)#
Compile this Module’s forward using
torch.compile()
.This Module’s __call__ method is compiled and all arguments are passed as-is to
torch.compile()
.See
torch.compile()
for details on the arguments for this function.
- cpu() T #
Moves all model parameters and buffers to the CPU.
Note
This method modifies the module in-place.
- Returns:
self
- Return type:
- cuda(device: Optional[Union[int, device]] = None) T #
Moves all model parameters and buffers to the GPU.
This also makes associated parameters and buffers different objects. So it should be called before constructing optimizer if the module will live on GPU while being optimized.
Note
This method modifies the module in-place.
- double() T #
Casts all floating point parameters and buffers to
double
datatype.Note
This method modifies the module in-place.
- Returns:
self
- Return type:
- eval() T #
Sets the module in evaluation mode.
This has any effect only on certain modules. See documentations of particular modules for details of their behaviors in training/evaluation mode, if they are affected, e.g.
Dropout
,BatchNorm
, etc.This is equivalent with
self.train(False)
.See Locally disabling gradient computation for a comparison between .eval() and several similar mechanisms that may be confused with it.
- Returns:
self
- Return type:
- extra_repr() str #
Set the extra representation of the module
To print customized extra information, you should re-implement this method in your own modules. Both single-line and multi-line strings are acceptable.
- float() T #
Casts all floating point parameters and buffers to
float
datatype.Note
This method modifies the module in-place.
- Returns:
self
- Return type:
- get_extra_state() Any #
Returns any extra state to include in the module’s state_dict. Implement this and a corresponding
set_extra_state()
for your module if you need to store extra state. This function is called when building the module’s state_dict().Note that extra state should be picklable to ensure working serialization of the state_dict. We only provide provide backwards compatibility guarantees for serializing Tensors; other objects may break backwards compatibility if their serialized pickled form changes.
- Returns:
Any extra state to store in the module’s state_dict
- Return type:
- get_submodule(target: str) Module #
Returns the submodule given by
target
if it exists, otherwise throws an error.For example, let’s say you have an
nn.Module
A
that looks like this:A( (net_b): Module( (net_c): Module( (conv): Conv2d(16, 33, kernel_size=(3, 3), stride=(2, 2)) ) (linear): Linear(in_features=100, out_features=200, bias=True) ) )
(The diagram shows an
nn.Module
A
.A
has a nested submodulenet_b
, which itself has two submodulesnet_c
andlinear
.net_c
then has a submoduleconv
.)To check whether or not we have the
linear
submodule, we would callget_submodule("net_b.linear")
. To check whether we have theconv
submodule, we would callget_submodule("net_b.net_c.conv")
.The runtime of
get_submodule
is bounded by the degree of module nesting intarget
. A query againstnamed_modules
achieves the same result, but it is O(N) in the number of transitive modules. So, for a simple check to see if some submodule exists,get_submodule
should always be used.- Parameters:
target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.)
- Returns:
The submodule referenced by
target
- Return type:
- Raises:
AttributeError – If the target string references an invalid path or resolves to something that is not an
nn.Module
- half() T #
Casts all floating point parameters and buffers to
half
datatype.Note
This method modifies the module in-place.
- Returns:
self
- Return type:
- ipu(device: Optional[Union[int, device]] = None) T #
Moves all model parameters and buffers to the IPU.
This also makes associated parameters and buffers different objects. So it should be called before constructing optimizer if the module will live on IPU while being optimized.
Note
This method modifies the module in-place.
- register_backward_hook(hook: Callable[[Module, Union[Tuple[Tensor, ...], Tensor], Union[Tuple[Tensor, ...], Tensor]], Union[None, Tuple[Tensor, ...], Tensor]]) RemovableHandle #
Registers a backward hook on the module.
This function is deprecated in favor of
register_full_backward_hook()
and the behavior of this function will change in future versions.- Returns:
a handle that can be used to remove the added hook by calling
handle.remove()
- Return type:
torch.utils.hooks.RemovableHandle
- register_forward_hook(hook: Union[Callable[[T, Tuple[Any, ...], Any], Optional[Any]], Callable[[T, Tuple[Any, ...], Dict[str, Any], Any], Optional[Any]]], *, prepend: bool = False, with_kwargs: bool = False, always_call: bool = False) RemovableHandle #
Registers a forward hook on the module.
The hook will be called every time after
forward()
has computed an output.If
with_kwargs
isFalse
or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to theforward
. The hook can modify the output. It can modify the input inplace but it will not have effect on forward since this is called afterforward()
is called. The hook should have the following signature:hook(module, args, output) -> None or modified output
If
with_kwargs
isTrue
, the forward hook will be passed thekwargs
given to the forward function and be expected to return the output possibly modified. The hook should have the following signature:hook(module, args, kwargs, output) -> None or modified output
- Parameters:
hook (Callable) – The user defined hook to be registered.
prepend (bool) – If
True
, the providedhook
will be fired before all existingforward
hooks on thistorch.nn.modules.Module
. Otherwise, the providedhook
will be fired after all existingforward
hooks on thistorch.nn.modules.Module
. Note that globalforward
hooks registered withregister_module_forward_hook()
will fire before all hooks registered by this method. Default:False
with_kwargs (bool) – If
True
, thehook
will be passed the kwargs given to the forward function. Default:False
always_call (bool) – If
True
thehook
will be run regardless of whether an exception is raised while calling the Module. Default:False
- Returns:
a handle that can be used to remove the added hook by calling
handle.remove()
- Return type:
torch.utils.hooks.RemovableHandle
- register_forward_pre_hook(hook: Union[Callable[[T, Tuple[Any, ...]], Optional[Any]], Callable[[T, Tuple[Any, ...], Dict[str, Any]], Optional[Tuple[Any, Dict[str, Any]]]]], *, prepend: bool = False, with_kwargs: bool = False) RemovableHandle #
Registers a forward pre-hook on the module.
The hook will be called every time before
forward()
is invoked.If
with_kwargs
is false or not specified, the input contains only the positional arguments given to the module. Keyword arguments won’t be passed to the hooks and only to theforward
. The hook can modify the input. User can either return a tuple or a single modified value in the hook. We will wrap the value into a tuple if a single value is returned (unless that value is already a tuple). The hook should have the following signature:hook(module, args) -> None or modified input
If
with_kwargs
is true, the forward pre-hook will be passed the kwargs given to the forward function. And if the hook modifies the input, both the args and kwargs should be returned. The hook should have the following signature:hook(module, args, kwargs) -> None or a tuple of modified input and kwargs
- Parameters:
hook (Callable) – The user defined hook to be registered.
prepend (bool) – If true, the provided
hook
will be fired before all existingforward_pre
hooks on thistorch.nn.modules.Module
. Otherwise, the providedhook
will be fired after all existingforward_pre
hooks on thistorch.nn.modules.Module
. Note that globalforward_pre
hooks registered withregister_module_forward_pre_hook()
will fire before all hooks registered by this method. Default:False
with_kwargs (bool) – If true, the
hook
will be passed the kwargs given to the forward function. Default:False
- Returns:
a handle that can be used to remove the added hook by calling
handle.remove()
- Return type:
torch.utils.hooks.RemovableHandle
- register_full_backward_hook(hook: Callable[[Module, Union[Tuple[Tensor, ...], Tensor], Union[Tuple[Tensor, ...], Tensor]], Union[None, Tuple[Tensor, ...], Tensor]], prepend: bool = False) RemovableHandle #
Registers a backward hook on the module.
The hook will be called every time the gradients with respect to a module are computed, i.e. the hook will execute if and only if the gradients with respect to module outputs are computed. The hook should have the following signature:
hook(module, grad_input, grad_output) -> tuple(Tensor) or None
The
grad_input
andgrad_output
are tuples that contain the gradients with respect to the inputs and outputs respectively. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the input that will be used in place ofgrad_input
in subsequent computations.grad_input
will only correspond to the inputs given as positional arguments and all kwarg arguments are ignored. Entries ingrad_input
andgrad_output
will beNone
for all non-Tensor arguments.For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function.
Warning
Modifying inputs or outputs inplace is not allowed when using backward hooks and will raise an error.
- Parameters:
hook (Callable) – The user-defined hook to be registered.
prepend (bool) – If true, the provided
hook
will be fired before all existingbackward
hooks on thistorch.nn.modules.Module
. Otherwise, the providedhook
will be fired after all existingbackward
hooks on thistorch.nn.modules.Module
. Note that globalbackward
hooks registered withregister_module_full_backward_hook()
will fire before all hooks registered by this method.
- Returns:
a handle that can be used to remove the added hook by calling
handle.remove()
- Return type:
torch.utils.hooks.RemovableHandle
- register_full_backward_pre_hook(hook: Callable[[Module, Union[Tuple[Tensor, ...], Tensor]], Union[None, Tuple[Tensor, ...], Tensor]], prepend: bool = False) RemovableHandle #
Registers a backward pre-hook on the module.
The hook will be called every time the gradients for the module are computed. The hook should have the following signature:
hook(module, grad_output) -> tuple[Tensor] or None
The
grad_output
is a tuple. The hook should not modify its arguments, but it can optionally return a new gradient with respect to the output that will be used in place ofgrad_output
in subsequent computations. Entries ingrad_output
will beNone
for all non-Tensor arguments.For technical reasons, when this hook is applied to a Module, its forward function will receive a view of each Tensor passed to the Module. Similarly the caller will receive a view of each Tensor returned by the Module’s forward function.
Warning
Modifying inputs inplace is not allowed when using backward hooks and will raise an error.
- Parameters:
hook (Callable) – The user-defined hook to be registered.
prepend (bool) – If true, the provided
hook
will be fired before all existingbackward_pre
hooks on thistorch.nn.modules.Module
. Otherwise, the providedhook
will be fired after all existingbackward_pre
hooks on thistorch.nn.modules.Module
. Note that globalbackward_pre
hooks registered withregister_module_full_backward_pre_hook()
will fire before all hooks registered by this method.
- Returns:
a handle that can be used to remove the added hook by calling
handle.remove()
- Return type:
torch.utils.hooks.RemovableHandle
- register_load_state_dict_post_hook(hook)#
Registers a post hook to be run after module’s
load_state_dict
is called.- It should have the following signature::
hook(module, incompatible_keys) -> None
The
module
argument is the current module that this hook is registered on, and theincompatible_keys
argument is aNamedTuple
consisting of attributesmissing_keys
andunexpected_keys
.missing_keys
is alist
ofstr
containing the missing keys andunexpected_keys
is alist
ofstr
containing the unexpected keys.The given incompatible_keys can be modified inplace if needed.
Note that the checks performed when calling
load_state_dict()
withstrict=True
are affected by modifications the hook makes tomissing_keys
orunexpected_keys
, as expected. Additions to either set of keys will result in an error being thrown whenstrict=True
, and clearing out both missing and unexpected keys will avoid an error.- Returns:
a handle that can be used to remove the added hook by calling
handle.remove()
- Return type:
torch.utils.hooks.RemovableHandle
- register_state_dict_pre_hook(hook)#
These hooks will be called with arguments:
self
,prefix
, andkeep_vars
before callingstate_dict
onself
. The registered hooks can be used to perform pre-processing before thestate_dict
call is made.
- requires_grad_(requires_grad: bool = True) T #
Change if autograd should record operations on parameters in this module.
This method sets the parameters’
requires_grad
attributes in-place.This method is helpful for freezing part of the module for finetuning or training parts of a model individually (e.g., GAN training).
See Locally disabling gradient computation for a comparison between .requires_grad_() and several similar mechanisms that may be confused with it.
- set_extra_state(state: Any)#
This function is called from
load_state_dict()
to handle any extra state found within the state_dict. Implement this function and a correspondingget_extra_state()
for your module if you need to store extra state within its state_dict.- Parameters:
state (dict) – Extra state from the state_dict
- to(*args, **kwargs)#
Moves and/or casts the parameters and buffers.
This can be called as
- to(device=None, dtype=None, non_blocking=False)
- to(dtype, non_blocking=False)
- to(tensor, non_blocking=False)
- to(memory_format=torch.channels_last)
Its signature is similar to
torch.Tensor.to()
, but only accepts floating point or complexdtype
s. In addition, this method will only cast the floating point or complex parameters and buffers todtype
(if given). The integral parameters and buffers will be moveddevice
, if that is given, but with dtypes unchanged. Whennon_blocking
is set, it tries to convert/move asynchronously with respect to the host if possible, e.g., moving CPU Tensors with pinned memory to CUDA devices.See below for examples.
Note
This method modifies the module in-place.
- Parameters:
device (
torch.device
) – the desired device of the parameters and buffers in this moduledtype (
torch.dtype
) – the desired floating point or complex dtype of the parameters and buffers in this moduletensor (torch.Tensor) – Tensor whose dtype and device are the desired dtype and device for all parameters and buffers in this module
memory_format (
torch.memory_format
) – the desired memory format for 4D parameters and buffers in this module (keyword only argument)
- Returns:
self
- Return type:
Examples:
>>> # xdoctest: +IGNORE_WANT("non-deterministic") >>> linear = nn.Linear(2, 2) >>> linear.weight Parameter containing: tensor([[ 0.1913, -0.3420], [-0.5113, -0.2325]]) >>> linear.to(torch.double) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1913, -0.3420], [-0.5113, -0.2325]], dtype=torch.float64) >>> # xdoctest: +REQUIRES(env:TORCH_DOCTEST_CUDA1) >>> gpu1 = torch.device("cuda:1") >>> linear.to(gpu1, dtype=torch.half, non_blocking=True) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1914, -0.3420], [-0.5112, -0.2324]], dtype=torch.float16, device='cuda:1') >>> cpu = torch.device("cpu") >>> linear.to(cpu) Linear(in_features=2, out_features=2, bias=True) >>> linear.weight Parameter containing: tensor([[ 0.1914, -0.3420], [-0.5112, -0.2324]], dtype=torch.float16) >>> linear = nn.Linear(2, 2, bias=None).to(torch.cdouble) >>> linear.weight Parameter containing: tensor([[ 0.3741+0.j, 0.2382+0.j], [ 0.5593+0.j, -0.4443+0.j]], dtype=torch.complex128) >>> linear(torch.ones(3, 2, dtype=torch.cdouble)) tensor([[0.6122+0.j, 0.1150+0.j], [0.6122+0.j, 0.1150+0.j], [0.6122+0.j, 0.1150+0.j]], dtype=torch.complex128)
- to_empty(*, device: Union[str, device], recurse: bool = True) T #
Moves the parameters and buffers to the specified device without copying storage.
- Parameters:
device (
torch.device
) – The desired device of the parameters and buffers in this module.recurse (bool) – Whether parameters and buffers of submodules should be recursively moved to the specified device.
- Returns:
self
- Return type:
- type(dst_type: Union[dtype, str]) T #
Casts all parameters and buffers to
dst_type
.Note
This method modifies the module in-place.
- xpu(device: Optional[Union[int, device]] = None) T #
Moves all model parameters and buffers to the XPU.
This also makes associated parameters and buffers different objects. So it should be called before constructing optimizer if the module will live on XPU while being optimized.
Note
This method modifies the module in-place.
- zero_grad(set_to_none: bool = True) None #
Resets gradients of all model parameters. See similar function under
torch.optim.Optimizer
for more context.- Parameters:
set_to_none (bool) – instead of setting to zero, set the grads to None. See
torch.optim.Optimizer.zero_grad()
for details.