菏泽网站建设费用,建设网站专业公司,用安卓做网站,西安模板建站网站分类目录#xff1a;《深入浅出Pytorch函数》总目录 Pytorch中所有网络的基类#xff0c;我们的模型也应该继承这个类。Modules也可以包含其它Modules,允许使用树结构嵌入他们#xff0c;我们还可以将子模块赋值给模型属性。
语法
torch.nn.Module(*args, **kwargs)方法
…分类目录《深入浅出Pytorch函数》总目录 Pytorch中所有网络的基类我们的模型也应该继承这个类。Modules也可以包含其它Modules,允许使用树结构嵌入他们我们还可以将子模块赋值给模型属性。
语法
torch.nn.Module(*args, **kwargs)方法
torch.nn.Module.apply
实例
import torch.nn as nn
import torch.nn.functional as Fclass Model(nn.Module):def __init__(self):super().__init__()self.conv1 nn.Conv2d(1, 20, 5) # submodule: Conv2dself.conv2 nn.Conv2d(20, 20, 5)def forward(self, x):x F.relu(self.conv1(x))return F.relu(self.conv2(x))通过上面方式赋值的submodule会被注册当调用.cuda() 的时候submodule的参数也会转换为cuda Tensor。
函数实现
from collections import OrderedDict, namedtuple
import itertools
import warnings
import functools
import weakrefimport torch
from ..parameter import Parameter
import torch.utils.hooks as hooksfrom torch import Tensor, device, dtype
from typing import Union, Tuple, Any, Callable, Iterator, Set, Optional, overload, TypeVar, Mapping, Dict, List
from ...utils.hooks import RemovableHandle__all__ [register_module_forward_pre_hook, register_module_forward_hook,register_module_full_backward_pre_hook, register_module_backward_hook,register_module_full_backward_hook, register_module_buffer_registration_hook,register_module_module_registration_hook, register_module_parameter_registration_hook, Module]_grad_t Union[Tuple[Tensor, ...], Tensor]
# See https://mypy.readthedocs.io/en/latest/generics.html#generic-methods-and-generic-self for the use
# of T to annotate self. Many methods of Module return self and we want those return values to be
# the type of the subclass, not the looser type of Module.
T TypeVar(T, boundModule)class _IncompatibleKeys(namedtuple(IncompatibleKeys, [missing_keys, unexpected_keys])):def __repr__(self):if not self.missing_keys and not self.unexpected_keys:return All keys matched successfullyreturn super().__repr__()__str__ __repr__def _addindent(s_, numSpaces):s s_.split(\n)# dont do anything for single-line stuffif len(s) 1:return s_first s.pop(0)s [(numSpaces * ) line for line in s]s \n.join(s)s first \n sreturn srThis tracks hooks common to all modules that are executed immediately before
.registering the buffer/module/parameter
_global_buffer_registration_hooks: Dict[int, Callable] OrderedDict()
_global_module_registration_hooks: Dict[int, Callable] OrderedDict()
_global_parameter_registration_hooks: Dict[int, Callable] OrderedDict()class _WrappedHook:def __init__(self, hook: Callable, module: Optional[Module] None):self.hook: Callable hookfunctools.update_wrapper(self, hook)self.with_module: bool Falseif module is not None:self.module: weakref.ReferenceType[Module] weakref.ref(module)self.with_module Truedef __call__(self, *args: Any, **kwargs: Any) - Any:if self.with_module:module self.module()if module is None:raise RuntimeError(You are trying to call the hook of a dead Module!)return self.hook(module, *args, **kwargs)return self.hook(*args, **kwargs)def __getstate__(self) - Dict:result {hook: self.hook, with_module: self.with_module}if self.with_module:result[module] self.module()return resultdef __setstate__(self, state: Dict):self.hook state[hook]self.with_module state[with_module]if self.with_module:if state[module] is None:raise RuntimeError(You are trying to revive the hook of a dead Module!)self.module weakref.ref(state[module])rThis tracks hooks common to all modules that are executed before/after
calling forward and backward. This is global state used for debugging/profiling
purposes
_global_backward_pre_hooks: Dict[int, Callable] OrderedDict()
_global_backward_hooks: Dict[int, Callable] OrderedDict()
_global_is_full_backward_hook: Optional[bool] None
_global_forward_pre_hooks: Dict[int, Callable] OrderedDict()
_global_forward_hooks: Dict[int, Callable] OrderedDict()_EXTRA_STATE_KEY_SUFFIX _extra_statedef register_module_buffer_registration_hook(hook: Callable[..., None]) - RemovableHandle:rRegisters a buffer registration hook common to all modules... warning ::This adds global state to the nn.Module moduleThe hook will be called every time :func:register_buffer is invoked.It should have the following signature::hook(module, name, buffer) - None or new bufferThe hook can modify the input or return a single modified value in the hook.Returns::class:torch.utils.hooks.RemovableHandle:a handle that can be used to remove the added hook by callinghandle.remove()handle hooks.RemovableHandle(_global_buffer_registration_hooks)_global_buffer_registration_hooks[handle.id] hookreturn handledef register_module_module_registration_hook(hook: Callable[..., None]) - RemovableHandle:rRegisters a module registration hook common to all modules... warning ::This adds global state to the nn.Module moduleThe hook will be called every time :func:register_module is invoked.It should have the following signature::hook(module, name, submodule) - None or new submoduleThe hook can modify the input or return a single modified value in the hook.Returns::class:torch.utils.hooks.RemovableHandle:a handle that can be used to remove the added hook by callinghandle.remove()handle hooks.RemovableHandle(_global_module_registration_hooks)_global_module_registration_hooks[handle.id] hookreturn handledef register_module_parameter_registration_hook(hook: Callable[..., None]) - RemovableHandle:rRegisters a parameter registration hook common to all modules... warning ::This adds global state to the nn.Module moduleThe hook will be called every time :func:register_parameter is invoked.It should have the following signature::hook(module, name, param) - None or new parameterThe hook can modify the input or return a single modified value in the hook.Returns::class:torch.utils.hooks.RemovableHandle:a handle that can be used to remove the added hook by callinghandle.remove()handle hooks.RemovableHandle(_global_parameter_registration_hooks)_global_parameter_registration_hooks[handle.id] hookreturn handledef register_module_forward_pre_hook(hook: Callable[..., None]) - RemovableHandle:rRegisters a forward pre-hook common to all modules... warning ::This adds global state to the nn.module moduleand it is only intended for debugging/profiling purposes.The hook will be called every time before :func:forward is invoked.It should have the following signature::hook(module, input) - None or modified inputThe input contains only the positional arguments given to the module.Keyword arguments wont be passed to the hooks and only to the forward.The hook can modify the input. User can either return a tuple or asingle modified value in the hook. We will wrap the value into a tupleif a single value is returned(unless that value is already a tuple).This hook has precedence over the specific module hooks registered withregister_forward_pre_hook.Returns::class:torch.utils.hooks.RemovableHandle:a handle that can be used to remove the added hook by callinghandle.remove()handle hooks.RemovableHandle(_global_forward_pre_hooks)_global_forward_pre_hooks[handle.id] hookreturn handledef register_module_forward_hook(hook: Callable[..., None]) - RemovableHandle:rRegisters a global forward hook for all the modules.. warning ::This adds global state to the nn.module moduleand it is only intended for debugging/profiling purposes.The hook will be called every time after :func:forward has computed an output.It should have the following signature::hook(module, input, output) - None or modified outputThe input contains only the positional arguments given to the module.Keyword arguments wont be passed to the hooks and only to the forward.The hook can modify the output. It can modify the input inplace butit will not have effect on forward since this is called after:func:forward is called.Returns::class:torch.utils.hooks.RemovableHandle:a handle that can be used to remove the added hook by callinghandle.remove()This hook will be executed before specific module hooks registered withregister_forward_hook.handle hooks.RemovableHandle(_global_forward_hooks)_global_forward_hooks[handle.id] hookreturn handledef register_module_backward_hook(hook: Callable[[Module, _grad_t, _grad_t], Union[None, _grad_t]]
) - RemovableHandle:rRegisters a backward hook common to all the modules.This function is deprecated in favor of:func:torch.nn.modules.module.register_module_full_backward_hookand the behavior of this function will change in future versions.Returns::class:torch.utils.hooks.RemovableHandle:a handle that can be used to remove the added hook by callinghandle.remove()global _global_is_full_backward_hookif _global_is_full_backward_hook is True:raise RuntimeError(Cannot use both regular backward hooks and full backward hooks as a global Module hook. Please use only one of them.)_global_is_full_backward_hook Falsehandle hooks.RemovableHandle(_global_backward_hooks)_global_backward_hooks[handle.id] hookreturn handledef register_module_full_backward_pre_hook(hook: Callable[[Module, _grad_t], Union[None, _grad_t]]
) - RemovableHandle:rRegisters a backward pre-hook common to all the modules... warning ::This adds global state to the nn.module moduleand it is only intended for debugging/profiling purposes.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) - Tensor or NoneThe :attr:grad_output is a tuple. The hook shouldnot modify its arguments, but it can optionally return a new gradient withrespect to the output that will be used in place of :attr:grad_output insubsequent computations. Entries in :attr:grad_output will be None forall non-Tensor arguments.For technical reasons, when this hook is applied to a Module, its forward function willreceive a view of each Tensor passed to the Module. Similarly the caller will receive a viewof each Tensor returned by the Modules forward function.Global hooks are called before hooks registered with register_backward_pre_hookReturns::class:torch.utils.hooks.RemovableHandle:a handle that can be used to remove the added hook by callinghandle.remove()handle hooks.RemovableHandle(_global_backward_pre_hooks)_global_backward_pre_hooks[handle.id] hookreturn handledef register_module_full_backward_hook(hook: Callable[[Module, _grad_t, _grad_t], Union[None, _grad_t]]
) - RemovableHandle:rRegisters a backward hook common to all the modules... warning ::This adds global state to the nn.module moduleand it is only intended for debugging/profiling purposes.The hook will be called every time the gradients with respect to a moduleare computed, i.e. the hook will execute if and only if the gradients withrespect to module outputs are computed. The hook should have the followingsignature::hook(module, grad_input, grad_output) - Tensor or NoneThe :attr:grad_input and :attr:grad_output are tuples. The hook shouldnot modify its arguments, but it can optionally return a new gradient withrespect to the input that will be used in place of :attr:grad_input insubsequent computations. :attr:grad_input will only correspond to the inputs givenas positional arguments and all kwarg arguments will not appear in the hook. Entriesin :attr:grad_input and :attr:grad_output will be None for all non-Tensorarguments.For technical reasons, when this hook is applied to a Module, its forward function willreceive a view of each Tensor passed to the Module. Similarly the caller will receive a viewof each Tensor returned by the Modules forward function.Global hooks are called before hooks registered with register_backward_hookReturns::class:torch.utils.hooks.RemovableHandle:a handle that can be used to remove the added hook by callinghandle.remove()global _global_is_full_backward_hookif _global_is_full_backward_hook is False:raise RuntimeError(Cannot use both regular backward hooks and full backward hooks as a global Module hook. Please use only one of them.)_global_is_full_backward_hook Truehandle hooks.RemovableHandle(_global_backward_hooks)_global_backward_hooks[handle.id] hookreturn handle# Trick mypy into not applying contravariance rules to inputs by defining
# forward as a value, rather than a function. See also
# https://github.com/python/mypy/issues/8795
def _forward_unimplemented(self, *input: Any) - None:rDefines the computation performed at every call.Should be overridden by all subclasses... note::Although the recipe for forward pass needs to be defined withinthis function, one should call the :class:Module instance afterwardsinstead of this since the former takes care of running theregistered hooks while the latter silently ignores them.raise NotImplementedError(fModule [{type(self).__name__}] is missing the required \forward\ function)class Module:rBase class for all neural network modules.Your models should also subclass this class.Modules can also contain other Modules, allowing to nest them ina tree structure. You can assign the submodules as regular attributes::import torch.nn as nnimport torch.nn.functional as Fclass Model(nn.Module):def __init__(self):super().__init__()self.conv1 nn.Conv2d(1, 20, 5)self.conv2 nn.Conv2d(20, 20, 5)def forward(self, x):x F.relu(self.conv1(x))return F.relu(self.conv2(x))Submodules assigned in this way will be registered, and will have theirparameters converted too when you call :meth:to, etc... note::As per the example above, an __init__() call to the parent classmust be made before assignment on the child.:ivar training: Boolean represents whether this module is in training orevaluation mode.:vartype training: booldump_patches: bool False_version: int 1rThis allows better BC support for :meth:load_state_dict. In:meth:state_dict, the version number will be saved as in the attribute_metadata of the returned state dict, and thus pickled. _metadata is adictionary with keys that follow the naming convention of state dict. See_load_from_state_dict on how to use this information in loading.If new parameters/buffers are added/removed from a module, this number shallbe bumped, and the modules _load_from_state_dict method can compare theversion number and do appropriate changes if the state dict is from beforethe change.training: bool_parameters: Dict[str, Optional[Parameter]]_buffers: Dict[str, Optional[Tensor]]_non_persistent_buffers_set: Set[str]_backward_pre_hooks: Dict[int, Callable]_backward_hooks: Dict[int, Callable]_is_full_backward_hook: Optional[bool]_forward_hooks: Dict[int, Callable]# Marks whether the corresponding _forward_hooks accept kwargs or not.# As JIT does not support Set[int], this dict is used as a set, where all# hooks represented in this dict accept kwargs._forward_hooks_with_kwargs: Dict[int, bool]_forward_pre_hooks: Dict[int, Callable]# Marks whether the corresponding _forward_hooks accept kwargs or not.# As JIT does not support Set[int], this dict is used as a set, where all# hooks represented in this dict accept kwargs._forward_pre_hooks_with_kwargs: Dict[int, bool]_state_dict_hooks: Dict[int, Callable]_load_state_dict_pre_hooks: Dict[int, Callable]_state_dict_pre_hooks: Dict[int, Callable]_load_state_dict_post_hooks: Dict[int, Callable]_modules: Dict[str, Optional[Module]]call_super_init: bool Falsedef __init__(self, *args, **kwargs) - None:Initializes internal Module state, shared by both nn.Module and ScriptModule.torch._C._log_api_usage_once(python.nn_module)# Backward compatibility: no args used to be allowed when call_super_initFalseif self.call_super_init is False and bool(kwargs):raise TypeError({}.__init__() got an unexpected keyword argument {}.format(type(self).__name__, next(iter(kwargs))))if self.call_super_init is False and bool(args):raise TypeError({}.__init__() takes 1 positional argument but {} were given.format(type(self).__name__, len(args) 1))Calls super().__setattr__(a, a) instead of the typical self.a ato avoid Module.__setattr__ overhead. Modules __setattr__ has specialhandling for parameters, submodules, and buffers but simply calls intosuper().__setattr__ for all other attributes.super().__setattr__(training, True)super().__setattr__(_parameters, OrderedDict())super().__setattr__(_buffers, OrderedDict())super().__setattr__(_non_persistent_buffers_set, set())super().__setattr__(_backward_pre_hooks, OrderedDict())super().__setattr__(_backward_hooks, OrderedDict())super().__setattr__(_is_full_backward_hook, None)super().__setattr__(_forward_hooks, OrderedDict())super().__setattr__(_forward_hooks_with_kwargs, OrderedDict())super().__setattr__(_forward_pre_hooks, OrderedDict())super().__setattr__(_forward_pre_hooks_with_kwargs, OrderedDict())super().__setattr__(_state_dict_hooks, OrderedDict())super().__setattr__(_state_dict_pre_hooks, OrderedDict())super().__setattr__(_load_state_dict_pre_hooks, OrderedDict())super().__setattr__(_load_state_dict_post_hooks, OrderedDict())super().__setattr__(_modules, OrderedDict())if self.call_super_init:super().__init__(*args, **kwargs)forward: Callable[..., Any] _forward_unimplementeddef register_buffer(self, name: str, tensor: Optional[Tensor], persistent: bool True) - None:rAdds a buffer to the module.This is typically used to register a buffer that should not to beconsidered a model parameter. For example, BatchNorms running_meanis not a parameter, but is part of the modules state. Buffers, bydefault, are persistent and will be saved alongside parameters. Thisbehavior can be changed by setting :attr:persistent to False. Theonly difference between a persistent buffer and a non-persistent bufferis that the latter will not be a part of this modules:attr:state_dict.Buffers can be accessed as attributes using given names.Args:name (str): name of the buffer. The buffer can be accessedfrom this module using the given nametensor (Tensor or None): buffer to be registered. If None, then operationsthat run on buffers, such as :attr:cuda, are ignored. If None,the buffer is **not** included in the modules :attr:state_dict.persistent (bool): whether the buffer is part of this modules:attr:state_dict.Example:: # xdoctest: SKIP(undefined vars) self.register_buffer(running_mean, torch.zeros(num_features))if persistent is False and isinstance(self, torch.jit.ScriptModule):raise RuntimeError(ScriptModule does not support non-persistent buffers)if _buffers not in self.__dict__:raise AttributeError(cannot assign buffer before Module.__init__() call)elif not isinstance(name, str):raise TypeError(buffer name should be a string. Got {}.format(torch.typename(name)))elif . in name:raise KeyError(buffer name cant contain \.\)elif name :raise KeyError(buffer name cant be empty string \\)elif hasattr(self, name) and name not in self._buffers:raise KeyError(attribute {} already exists.format(name))elif tensor is not None and not isinstance(tensor, torch.Tensor):raise TypeError(cannot assign {} object to buffer {} (torch Tensor or None required).format(torch.typename(tensor), name))else:for hook in _global_buffer_registration_hooks.values():output hook(self, name, tensor)if output is not None:tensor outputself._buffers[name] tensorif persistent:self._non_persistent_buffers_set.discard(name)else:self._non_persistent_buffers_set.add(name)def register_parameter(self, name: str, param: Optional[Parameter]) - None:rAdds a parameter to the module.The parameter can be accessed as an attribute using given name.Args:name (str): name of the parameter. The parameter can be accessedfrom this module using the given nameparam (Parameter or None): parameter to be added to the module. IfNone, then operations that run on parameters, such as :attr:cuda,are ignored. If None, the parameter is **not** included in themodules :attr:state_dict.if _parameters not in self.__dict__:raise AttributeError(cannot assign parameter before Module.__init__() call)elif not isinstance(name, str):raise TypeError(parameter name should be a string. Got {}.format(torch.typename(name)))elif . in name:raise KeyError(parameter name cant contain \.\)elif name :raise KeyError(parameter name cant be empty string \\)elif hasattr(self, name) and name not in self._parameters:raise KeyError(attribute {} already exists.format(name))if param is None:self._parameters[name] Noneelif not isinstance(param, Parameter):raise TypeError(cannot assign {} object to parameter {} (torch.nn.Parameter or None required).format(torch.typename(param), name))elif param.grad_fn:raise ValueError(Cannot assign non-leaf Tensor to parameter {0}. Model parameters must be created explicitly. To express {0} as a function of another Tensor, compute the value in the forward() method..format(name))else:for hook in _global_parameter_registration_hooks.values():output hook(self, name, param)if output is not None:param outputself._parameters[name] paramdef add_module(self, name: str, module: Optional[Module]) - None:rAdds a child module to the current module.The module can be accessed as an attribute using the given name.Args:name (str): name of the child module. The child module can beaccessed from this module using the given namemodule (Module): child module to be added to the module.if not isinstance(module, Module) and module is not None:raise TypeError({} is not a Module subclass.format(torch.typename(module)))elif not isinstance(name, str):raise TypeError(module name should be a string. Got {}.format(torch.typename(name)))elif hasattr(self, name) and name not in self._modules:raise KeyError(attribute {} already exists.format(name))elif . in name:raise KeyError(module name cant contain \.\, got: {}.format(name))elif name :raise KeyError(module name cant be empty string \\)for hook in _global_module_registration_hooks.values():output hook(self, name, module)if output is not None:module outputself._modules[name] moduledef register_module(self, name: str, module: Optional[Module]) - None:rAlias for :func:add_module.self.add_module(name, module)def get_submodule(self, target: str) - Module:Returns the submodule given by target if it exists,otherwise throws an error.For example, lets say you have an nn.Module A thatlooks like this:.. code-block:: textA((net_b): Module((net_c): Module((conv): Conv2d(16, 33, kernel_size(3, 3), stride(2, 2)))(linear): Linear(in_features100, out_features200, biasTrue)))(The diagram shows an nn.Module A. A has a nestedsubmodule net_b, which itself has two submodules net_cand linear. net_c then has a submodule conv.)To check whether or not we have the linear submodule, wewould call get_submodule(net_b.linear). To check whetherwe have the conv submodule, we would callget_submodule(net_b.net_c.conv).The runtime of get_submodule is bounded by the degreeof module nesting in target. A query againstnamed_modules achieves the same result, but it is O(N) inthe number of transitive modules. So, for a simple check to seeif some submodule exists, get_submodule should always beused.Args:target: The fully-qualified string name of the submoduleto look for. (See above example for how to specify afully-qualified string.)Returns:torch.nn.Module: The submodule referenced by targetRaises:AttributeError: If the target string references an invalidpath or resolves to something that is not annn.Moduleif target :return selfatoms: List[str] target.split(.)mod: torch.nn.Module selffor item in atoms:if not hasattr(mod, item):raise AttributeError(mod._get_name() has no attribute item )mod getattr(mod, item)if not isinstance(mod, torch.nn.Module):raise AttributeError( item is not an nn.Module)return moddef get_parameter(self, target: str) - Parameter:Returns the parameter given by target if it exists,otherwise throws an error.See the docstring for get_submodule for a more detailedexplanation of this methods functionality as well as how tocorrectly specify target.Args:target: The fully-qualified string name of the Parameterto look for. (See get_submodule for how to specify afully-qualified string.)Returns:torch.nn.Parameter: The Parameter referenced by targetRaises:AttributeError: If the target string references an invalidpath or resolves to something that is not annn.Parametermodule_path, _, param_name target.rpartition(.)mod: torch.nn.Module self.get_submodule(module_path)if not hasattr(mod, param_name):raise AttributeError(mod._get_name() has no attribute param_name )param: torch.nn.Parameter getattr(mod, param_name)if not isinstance(param, torch.nn.Parameter):raise AttributeError( param_name is not an nn.Parameter)return paramdef get_buffer(self, target: str) - Tensor:Returns the buffer given by target if it exists,otherwise throws an error.See the docstring for get_submodule for a more detailedexplanation of this methods functionality as well as how tocorrectly specify target.Args:target: The fully-qualified string name of the bufferto look for. (See get_submodule for how to specify afully-qualified string.)Returns:torch.Tensor: The buffer referenced by targetRaises:AttributeError: If the target string references an invalidpath or resolves to something that is not abuffermodule_path, _, buffer_name target.rpartition(.)mod: torch.nn.Module self.get_submodule(module_path)if not hasattr(mod, buffer_name):raise AttributeError(mod._get_name() has no attribute buffer_name )buffer: torch.Tensor getattr(mod, buffer_name)if buffer_name not in mod._buffers:raise AttributeError( buffer_name is not a buffer)return bufferdef get_extra_state(self) - Any:Returns any extra state to include in the modules state_dict.Implement this and a corresponding :func:set_extra_state for your moduleif you need to store extra state. This function is called when building themodules state_dict().Note that extra state should be picklable to ensure working serializationof the state_dict. We only provide provide backwards compatibility guaranteesfor serializing Tensors; other objects may break backwards compatibility iftheir serialized pickled form changes.Returns:object: Any extra state to store in the modules state_dictraise RuntimeError(Reached a code path in Module.get_extra_state() that should never be called. Please file an issue at https://github.com/pytorch/pytorch/issues/new?templatebug-report.yml to report this bug.)def set_extra_state(self, state: Any):This function is called from :func:load_state_dict to handle any extra statefound within the state_dict. Implement this function and a corresponding:func:get_extra_state for your module if you need to store extra state within itsstate_dict.Args:state (dict): Extra state from the state_dictraise RuntimeError(Reached a code path in Module.set_extra_state() that should never be called. Please file an issue at https://github.com/pytorch/pytorch/issues/new?templatebug-report.yml to report this bug.)def _apply(self, fn):for module in self.children():module._apply(fn)def compute_should_use_set_data(tensor, tensor_applied):if torch._has_compatible_shallow_copy_type(tensor, tensor_applied):# If the new tensor has compatible tensor type as the existing tensor,# the current behavior is to change the tensor in-place using .data ,# and the future behavior is to overwrite the existing tensor. However,# changing the current behavior is a BC-breaking change, and we want it# to happen in future releases. So for now we introduce the# torch.__future__.get_overwrite_module_params_on_conversion()# global flag to let the user control whether they want the future# behavior of overwriting the existing tensor or not.return not torch.__future__.get_overwrite_module_params_on_conversion()else:return Falsefor key, param in self._parameters.items():if param is None:continue# Tensors stored in modules are graph leaves, and we dont want to# track autograd history of param_applied, so we have to use# with torch.no_grad():with torch.no_grad():param_applied fn(param)should_use_set_data compute_should_use_set_data(param, param_applied)if should_use_set_data:param.data param_appliedout_param paramelse:assert isinstance(param, Parameter)assert param.is_leafout_param Parameter(param_applied, param.requires_grad)self._parameters[key] out_paramif param.grad is not None:with torch.no_grad():grad_applied fn(param.grad)should_use_set_data compute_should_use_set_data(param.grad, grad_applied)if should_use_set_data:assert out_param.grad is not Noneout_param.grad.data grad_appliedelse:assert param.grad.is_leafout_param.grad grad_applied.requires_grad_(param.grad.requires_grad)for key, buf in self._buffers.items():if buf is not None:self._buffers[key] fn(buf)return selfdef apply(self: T, fn: Callable[[Module], None]) - T:rApplies fn recursively to every submodule (as returned by .children())as well as self. Typical use includes initializing the parameters of a model(see also :ref:nn-init-doc).Args:fn (:class:Module - None): function to be applied to each submoduleReturns:Module: selfExample:: torch.no_grad() def init_weights(m): print(m) if type(m) nn.Linear: m.weight.fill_(1.0) print(m.weight) net nn.Sequential(nn.Linear(2, 2), nn.Linear(2, 2)) net.apply(init_weights)Linear(in_features2, out_features2, biasTrue)Parameter containing:tensor([[1., 1.],[1., 1.]], requires_gradTrue)Linear(in_features2, out_features2, biasTrue)Parameter containing:tensor([[1., 1.],[1., 1.]], requires_gradTrue)Sequential((0): Linear(in_features2, out_features2, biasTrue)(1): Linear(in_features2, out_features2, biasTrue))for module in self.children():module.apply(fn)fn(self)return selfdef cuda(self: T, device: Optional[Union[int, device]] None) - T:rMoves all model parameters and buffers to the GPU.This also makes associated parameters and buffers different objects. Soit should be called before constructing optimizer if the module willlive on GPU while being optimized... note::This method modifies the module in-place.Args:device (int, optional): if specified, all parameters will becopied to that deviceReturns:Module: selfreturn self._apply(lambda t: t.cuda(device))def ipu(self: T, device: Optional[Union[int, device]] None) - T:rMoves all model parameters and buffers to the IPU.This also makes associated parameters and buffers different objects. Soit should be called before constructing optimizer if the module willlive on IPU while being optimized... note::This method modifies the module in-place.Arguments:device (int, optional): if specified, all parameters will becopied to that deviceReturns:Module: selfreturn self._apply(lambda t: t.ipu(device))def xpu(self: T, device: Optional[Union[int, device]] None) - T:rMoves all model parameters and buffers to the XPU.This also makes associated parameters and buffers different objects. Soit should be called before constructing optimizer if the module willlive on XPU while being optimized... note::This method modifies the module in-place.Arguments:device (int, optional): if specified, all parameters will becopied to that deviceReturns:Module: selfreturn self._apply(lambda t: t.xpu(device))def cpu(self: T) - T:rMoves all model parameters and buffers to the CPU... note::This method modifies the module in-place.Returns:Module: selfreturn self._apply(lambda t: t.cpu())def type(self: T, dst_type: Union[dtype, str]) - T:rCasts all parameters and buffers to :attr:dst_type... note::This method modifies the module in-place.Args:dst_type (type or string): the desired typeReturns:Module: selfreturn self._apply(lambda t: t.type(dst_type))def float(self: T) - T:rCasts all floating point parameters and buffers to float datatype... note::This method modifies the module in-place.Returns:Module: selfreturn self._apply(lambda t: t.float() if t.is_floating_point() else t)def double(self: T) - T:rCasts all floating point parameters and buffers to double datatype... note::This method modifies the module in-place.Returns:Module: selfreturn self._apply(lambda t: t.double() if t.is_floating_point() else t)def half(self: T) - T:rCasts all floating point parameters and buffers to half datatype... note::This method modifies the module in-place.Returns:Module: selfreturn self._apply(lambda t: t.half() if t.is_floating_point() else t)def bfloat16(self: T) - T:rCasts all floating point parameters and buffers to bfloat16 datatype... note::This method modifies the module in-place.Returns:Module: selfreturn self._apply(lambda t: t.bfloat16() if t.is_floating_point() else t)def to_empty(self: T, *, device: Union[str, device]) - T:rMoves the parameters and buffers to the specified device without copying storage.Args:device (:class:torch.device): The desired device of the parametersand buffers in this module.Returns:Module: selfreturn self._apply(lambda t: torch.empty_like(t, devicedevice))overloaddef to(self: T, device: Optional[Union[int, device]] ..., dtype: Optional[Union[dtype, str]] ...,non_blocking: bool ...) - T:...overloaddef to(self: T, dtype: Union[dtype, str], non_blocking: bool ...) - T:...overloaddef to(self: T, tensor: Tensor, non_blocking: bool ...) - T:...def to(self, *args, **kwargs):rMoves and/or casts the parameters and buffers.This can be called as.. function:: to(deviceNone, dtypeNone, non_blockingFalse):noindex:.. function:: to(dtype, non_blockingFalse):noindex:.. function:: to(tensor, non_blockingFalse):noindex:.. function:: to(memory_formattorch.channels_last):noindex:Its signature is similar to :meth:torch.Tensor.to, but only acceptsfloating point or complex :attr:dtype\ s. In addition, this method willonly cast the floating point or complex parameters and buffers to :attr:dtype(if given). The integral parameters and buffers will be moved:attr:device, if that is given, but with dtypes unchanged. When:attr:non_blocking is set, it tries to convert/move asynchronouslywith respect to the host if possible, e.g., moving CPU Tensors withpinned memory to CUDA devices.See below for examples... note::This method modifies the module in-place.Args:device (:class:torch.device): the desired device of the parametersand buffers in this moduledtype (:class:torch.dtype): the desired floating point or complex dtype ofthe parameters and buffers in this moduletensor (torch.Tensor): Tensor whose dtype and device are the desireddtype and device for all parameters and buffers in this modulememory_format (:class:torch.memory_format): the desired memoryformat for 4D parameters and buffers in this module (keywordonly argument)Returns:Module: selfExamples:: # xdoctest: IGNORE_WANT(non-deterministic) linear nn.Linear(2, 2) linear.weightParameter containing:tensor([[ 0.1913, -0.3420],[-0.5113, -0.2325]]) linear.to(torch.double)Linear(in_features2, out_features2, biasTrue) linear.weightParameter containing:tensor([[ 0.1913, -0.3420],[-0.5113, -0.2325]], dtypetorch.float64) # xdoctest: REQUIRES(env:TORCH_DOCTEST_CUDA1) gpu1 torch.device(cuda:1) linear.to(gpu1, dtypetorch.half, non_blockingTrue)Linear(in_features2, out_features2, biasTrue) linear.weightParameter containing:tensor([[ 0.1914, -0.3420],[-0.5112, -0.2324]], dtypetorch.float16, devicecuda:1) cpu torch.device(cpu) linear.to(cpu)Linear(in_features2, out_features2, biasTrue) linear.weightParameter containing:tensor([[ 0.1914, -0.3420],[-0.5112, -0.2324]], dtypetorch.float16) linear nn.Linear(2, 2, biasNone).to(torch.cdouble) linear.weightParameter containing:tensor([[ 0.37410.j, 0.23820.j],[ 0.55930.j, -0.44430.j]], dtypetorch.complex128) linear(torch.ones(3, 2, dtypetorch.cdouble))tensor([[0.61220.j, 0.11500.j],[0.61220.j, 0.11500.j],[0.61220.j, 0.11500.j]], dtypetorch.complex128)device, dtype, non_blocking, convert_to_format torch._C._nn._parse_to(*args, **kwargs)if dtype is not None:if not (dtype.is_floating_point or dtype.is_complex):raise TypeError(nn.Module.to only accepts floating point or complex dtypes, but got desired dtype{}.format(dtype))if dtype.is_complex:warnings.warn(Complex modules are a new feature under active development whose design may change, and some modules might not work as expected when using complex tensors as parameters or buffers. Please file an issue at https://github.com/pytorch/pytorch/issues/new?templatebug-report.yml if a complex module does not work as expected.)def convert(t):if convert_to_format is not None and t.dim() in (4, 5):return t.to(device, dtype if t.is_floating_point() or t.is_complex() else None,non_blocking, memory_formatconvert_to_format)return t.to(device, dtype if t.is_floating_point() or t.is_complex() else None, non_blocking)return self._apply(convert)def register_full_backward_pre_hook(self,hook: Callable[[Module, _grad_t], Union[None, _grad_t]],prepend: bool False,) - RemovableHandle:rRegisters 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) - Tensor or NoneThe :attr:grad_output is a tuple. The hook shouldnot modify its arguments, but it can optionally return a new gradient withrespect to the output that will be used in place of :attr:grad_output insubsequent computations. Entries in :attr:grad_output will be None forall non-Tensor arguments.For technical reasons, when this hook is applied to a Module, its forward function willreceive a view of each Tensor passed to the Module. Similarly the caller will receive a viewof each Tensor returned by the Modules forward function... warning ::Modifying inputs inplace is not allowed when using backward hooks andwill raise an error.Args:hook (Callable): The user-defined hook to be registered.prepend (bool): If true, the provided hook will be fired beforeall existing backward_pre hooks on this:class:torch.nn.modules.Module. Otherwise, the providedhook will be fired after all existing backward_pre hookson this :class:torch.nn.modules.Module. Note that globalbackward_pre hooks registered with:func:register_module_full_backward_pre_hook will fire beforeall hooks registered by this method.Returns::class:torch.utils.hooks.RemovableHandle:a handle that can be used to remove the added hook by callinghandle.remove()handle hooks.RemovableHandle(self._backward_pre_hooks)self._backward_pre_hooks[handle.id] hookif prepend:self._backward_pre_hooks.move_to_end(handle.id, lastFalse) # type: ignore[attr-defined]return handledef register_backward_hook(self, hook: Callable[[Module, _grad_t, _grad_t], Union[None, _grad_t]]) - RemovableHandle:rRegisters a backward hook on the module.This function is deprecated in favor of :meth:~torch.nn.Module.register_full_backward_hook andthe behavior of this function will change in future versions.Returns::class:torch.utils.hooks.RemovableHandle:a handle that can be used to remove the added hook by callinghandle.remove()if self._is_full_backward_hook is True:raise RuntimeError(Cannot use both regular backward hooks and full backward hooks on a single Module. Please use only one of them.)self._is_full_backward_hook Falsehandle hooks.RemovableHandle(self._backward_hooks)self._backward_hooks[handle.id] hookreturn handledef register_full_backward_hook(self,hook: Callable[[Module, _grad_t, _grad_t], Union[None, _grad_t]],prepend: bool False,) - RemovableHandle:rRegisters a backward hook on the module.The hook will be called every time the gradients with respect to a moduleare computed, i.e. the hook will execute if and only if the gradients withrespect to module outputs are computed. The hook should have the followingsignature::hook(module, grad_input, grad_output) - tuple(Tensor) or NoneThe :attr:grad_input and :attr:grad_output are tuples that contain the gradientswith respect to the inputs and outputs respectively. The hook shouldnot modify its arguments, but it can optionally return a new gradient withrespect to the input that will be used in place of :attr:grad_input insubsequent computations. :attr:grad_input will only correspond to the inputs givenas positional arguments and all kwarg arguments are ignored. Entriesin :attr:grad_input and :attr:grad_output will be None for all non-Tensorarguments.For technical reasons, when this hook is applied to a Module, its forward function willreceive a view of each Tensor passed to the Module. Similarly the caller will receive a viewof each Tensor returned by the Modules forward function... warning ::Modifying inputs or outputs inplace is not allowed when using backward hooks andwill raise an error.Args:hook (Callable): The user-defined hook to be registered.prepend (bool): If true, the provided hook will be fired beforeall existing backward hooks on this:class:torch.nn.modules.Module. Otherwise, the providedhook will be fired after all existing backward hooks onthis :class:torch.nn.modules.Module. Note that globalbackward hooks registered with:func:register_module_full_backward_hook will fire beforeall hooks registered by this method.Returns::class:torch.utils.hooks.RemovableHandle:a handle that can be used to remove the added hook by callinghandle.remove()if self._is_full_backward_hook is False:raise RuntimeError(Cannot use both regular backward hooks and full backward hooks on a single Module. Please use only one of them.)self._is_full_backward_hook Truehandle hooks.RemovableHandle(self._backward_hooks)self._backward_hooks[handle.id] hookif prepend:self._backward_hooks.move_to_end(handle.id, lastFalse) # type: ignore[attr-defined]return handledef _get_backward_hooks(self):rReturns the backward hooks for use in the call function.It returns two lists, one with the full backward hooks and one with the non-fullbackward hooks.full_backward_hooks: List[Callable] []if (_global_is_full_backward_hook is True):full_backward_hooks _global_backward_hooks.values()if (self._is_full_backward_hook is True):full_backward_hooks self._backward_hooks.values()non_full_backward_hooks: List[Callable] []if (_global_is_full_backward_hook is False):non_full_backward_hooks _global_backward_hooks.values()if (self._is_full_backward_hook is False):non_full_backward_hooks self._backward_hooks.values()return full_backward_hooks, non_full_backward_hooksdef _get_backward_pre_hooks(self):backward_pre_hooks: List[Callable] []backward_pre_hooks _global_backward_pre_hooks.values()backward_pre_hooks self._backward_pre_hooks.values()return backward_pre_hooksdef _maybe_warn_non_full_backward_hook(self, inputs, result, grad_fn):if not isinstance(result, torch.Tensor):if not (isinstance(result, tuple) and all(isinstance(r, torch.Tensor) for r in result)):warnings.warn(Using non-full backward hooks on a Module that does not return a single Tensor or a tuple of Tensors is deprecated and will be removed in future versions. This hook will be missing some of the grad_output. Please use register_full_backward_hook to get the documented behavior.)returnelse:result (result,)if not isinstance(inputs, torch.Tensor):if not (isinstance(inputs, tuple) and all(isinstance(i, torch.Tensor) for i in inputs)):warnings.warn(Using non-full backward hooks on a Module that does not take as input a single Tensor or a tuple of Tensors is deprecated and will be removed in future versions. This hook will be missing some of the grad_input. Please use register_full_backward_hook to get the documented behavior.)returnelse:inputs (inputs,)# At this point we are sure that inputs and result are tuple of Tensorsout_grad_fn {r.grad_fn for r in result if r.grad_fn is not None}if len(out_grad_fn) 0 or (len(out_grad_fn) 1 and grad_fn not in out_grad_fn):warnings.warn(Using a non-full backward hook when outputs are nested in python data structure is deprecated and will be removed in future versions. This hook will be missing some grad_output.)elif len(out_grad_fn) 1:warnings.warn(Using a non-full backward hook when outputs are generated by different autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_output. Please use register_full_backward_hook to get the documented behavior.)else:# At this point the grad_ouput part of the hook will most likely be correctinputs_grad_fn {i.grad_fn for i in inputs if i.grad_fn is not None}next_functions {n[0] for n in grad_fn.next_functions}if inputs_grad_fn ! next_functions:warnings.warn(Using a non-full backward hook when the forward contains multiple autograd Nodes is deprecated and will be removed in future versions. This hook will be missing some grad_input. Please use register_full_backward_hook to get the documented behavior.)def register_forward_pre_hook(self,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:rRegisters a forward pre-hook on the module.The hook will be called every time before :func:forward is invoked.If with_kwargs is false or not specified, the input contains onlythe positional arguments given to the module. Keyword arguments wont bepassed to the hooks and only to the forward. The hook can modify theinput. User can either return a tuple or a single modified value in thehook. 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 thefollowing signature::hook(module, args) - None or modified inputIf with_kwargs is true, the forward pre-hook will be passed thekwargs given to the forward function. And if the hook modifies theinput, both the args and kwargs should be returned. The hook should havethe following signature::hook(module, args, kwargs) - None or a tuple of modified input and kwargsArgs:hook (Callable): The user defined hook to be registered.prepend (bool): If true, the provided hook will be fired beforeall existing forward_pre hooks on this:class:torch.nn.modules.Module. Otherwise, the providedhook will be fired after all existing forward_pre hookson this :class:torch.nn.modules.Module. Note that globalforward_pre hooks registered with:func:register_module_forward_pre_hook will fire before allhooks registered by this method.Default: Falsewith_kwargs (bool): If true, the hook will be passed the kwargsgiven to the forward function.Default: FalseReturns::class:torch.utils.hooks.RemovableHandle:a handle that can be used to remove the added hook by callinghandle.remove()handle hooks.RemovableHandle(self._forward_pre_hooks,extra_dictself._forward_pre_hooks_with_kwargs)self._forward_pre_hooks[handle.id] hookif with_kwargs:self._forward_pre_hooks_with_kwargs[handle.id] Trueif prepend:self._forward_pre_hooks.move_to_end(handle.id, lastFalse) # type: ignore[attr-defined]return handledef register_forward_hook(self,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,) - RemovableHandle:rRegisters a forward hook on the module.The hook will be called every time after :func:forward has computed an output.If with_kwargs is False or not specified, the input contains onlythe positional arguments given to the module. Keyword arguments wont bepassed to the hooks and only to the forward. The hook can modify theoutput. It can modify the input inplace but it will not have effect onforward since this is called after :func:forward is called. The hookshould have the following signature::hook(module, args, output) - None or modified outputIf with_kwargs is True, the forward hook will be passed thekwargs given to the forward function and be expected to return theoutput possibly modified. The hook should have the following signature::hook(module, args, kwargs, output) - None or modified outputArgs:hook (Callable): The user defined hook to be registered.prepend (bool): If True, the provided hook will be firedbefore all existing forward hooks on this:class:torch.nn.modules.Module. Otherwise, the providedhook will be fired after all existing forward hooks onthis :class:torch.nn.modules.Module. Note that globalforward hooks registered with:func:register_module_forward_hook will fire before all hooksregistered by this method.Default: Falsewith_kwargs (bool): If True, the hook will be passed thekwargs given to the forward function.Default: FalseReturns::class:torch.utils.hooks.RemovableHandle:a handle that can be used to remove the added hook by callinghandle.remove()handle hooks.RemovableHandle(self._forward_hooks,extra_dictself._forward_hooks_with_kwargs)self._forward_hooks[handle.id] hookif with_kwargs:self._forward_hooks_with_kwargs[handle.id] Trueif prepend:self._forward_hooks.move_to_end(handle.id, lastFalse) # type: ignore[attr-defined]return handledef _slow_forward(self, *input, **kwargs):tracing_state torch._C._get_tracing_state()if not tracing_state or isinstance(self.forward, torch._C.ScriptMethod):return self.forward(*input, **kwargs)recording_scopes torch.jit._trace._trace_module_map is not Noneif recording_scopes:# type ignore was added because at this point one knows that# torch.jit._trace._trace_module_map is not Optional and has type Dict[Any, Any]name torch.jit._trace._trace_module_map[self] if self in torch.jit._trace._trace_module_map else None # type: ignore[index, operator] # noqa: B950if name:tracing_state.push_scope(name)else:recording_scopes Falsetry:result self.forward(*input, **kwargs)finally:if recording_scopes:tracing_state.pop_scope()return resultdef _call_impl(self, *args, **kwargs):forward_call (self._slow_forward if torch._C._get_tracing_state() else self.forward)# If we dont have any hooks, we want to skip the rest of the logic in# this function, and just call forward.if not (self._backward_hooks or self._backward_pre_hooks or self._forward_hooks or self._forward_pre_hooksor _global_backward_pre_hooks or _global_backward_hooksor _global_forward_hooks or _global_forward_pre_hooks):return forward_call(*args, **kwargs)# Do not call functions when jit is usedfull_backward_hooks, non_full_backward_hooks [], []backward_pre_hooks []if self._backward_pre_hooks or _global_backward_pre_hooks:backward_pre_hooks self._get_backward_pre_hooks()if self._backward_hooks or _global_backward_hooks:full_backward_hooks, non_full_backward_hooks self._get_backward_hooks()if _global_forward_pre_hooks or self._forward_pre_hooks:for hook_id, hook in (*_global_forward_pre_hooks.items(),*self._forward_pre_hooks.items(),):if hook_id in self._forward_pre_hooks_with_kwargs:result hook(self, args, kwargs) # type: ignore[misc]if result is not None:if isinstance(result, tuple) and len(result) 2:args, kwargs resultelse:raise RuntimeError(forward pre-hook must return None or a tuple fof (new_args, new_kwargs), but got {result}.)else:result hook(self, args)if result is not None:if not isinstance(result, tuple):result (result,)args resultbw_hook Noneif full_backward_hooks or backward_pre_hooks:bw_hook hooks.BackwardHook(self, full_backward_hooks, backward_pre_hooks)args bw_hook.setup_input_hook(args)result forward_call(*args, **kwargs)if _global_forward_hooks or self._forward_hooks:for hook_id, hook in (*_global_forward_hooks.items(),*self._forward_hooks.items(),):if hook_id in self._forward_hooks_with_kwargs:hook_result hook(self, args, kwargs, result)else:hook_result hook(self, args, result)if hook_result is not None:result hook_resultif bw_hook:if not isinstance(result, (torch.Tensor, tuple)):warnings.warn(For backward hooks to be called, module output should be a Tensor or a tuple of Tensorsf but received {type(result)})result bw_hook.setup_output_hook(result)# Handle the non-full backward hooksif non_full_backward_hooks:var resultwhile not isinstance(var, torch.Tensor):if isinstance(var, dict):var next((v for v in var.values() if isinstance(v, torch.Tensor)))else:var var[0]grad_fn var.grad_fnif grad_fn is not None:for hook in non_full_backward_hooks:grad_fn.register_hook(_WrappedHook(hook, self))self._maybe_warn_non_full_backward_hook(args, result, grad_fn)return result__call__ : Callable[..., Any] _call_impldef __setstate__(self, state):self.__dict__.update(state)# Support loading old checkpoints that dont have the following attrs:if _forward_pre_hooks not in self.__dict__:self._forward_pre_hooks OrderedDict()if _forward_pre_hooks_with_kwargs not in self.__dict__:self._forward_pre_hooks_with_kwargs OrderedDict()if _forward_hooks_with_kwargs not in self.__dict__:self._forward_hooks_with_kwargs OrderedDict()if _state_dict_hooks not in self.__dict__:self._state_dict_hooks OrderedDict()if _state_dict_pre_hooks not in self.__dict__:self._state_dict_pre_hooks OrderedDict()if _load_state_dict_pre_hooks not in self.__dict__:self._load_state_dict_pre_hooks OrderedDict()if _load_state_dict_post_hooks not in self.__dict__:self._load_state_dict_post_hooks OrderedDict()if _non_persistent_buffers_set not in self.__dict__:self._non_persistent_buffers_set set()if _is_full_backward_hook not in self.__dict__:self._is_full_backward_hook Noneif _backward_pre_hooks not in self.__dict__:self._backward_pre_hooks OrderedDict()def __getattr__(self, name: str) - Union[Tensor, Module]:if _parameters in self.__dict__:_parameters self.__dict__[_parameters]if name in _parameters:return _parameters[name]if _buffers in self.__dict__:_buffers self.__dict__[_buffers]if name in _buffers:return _buffers[name]if _modules in self.__dict__:modules self.__dict__[_modules]if name in modules:return modules[name]raise AttributeError({} object has no attribute {}.format(type(self).__name__, name))def __setattr__(self, name: str, value: Union[Tensor, Module]) - None:def remove_from(*dicts_or_sets):for d in dicts_or_sets:if name in d:if isinstance(d, dict):del d[name]else:d.discard(name)params self.__dict__.get(_parameters)if isinstance(value, Parameter):if params is None:raise AttributeError(cannot assign parameters before Module.__init__() call)remove_from(self.__dict__, self._buffers, self._modules, self._non_persistent_buffers_set)self.register_parameter(name, value)elif params is not None and name in params:if value is not None:raise TypeError(cannot assign {} as parameter {} (torch.nn.Parameter or None expected).format(torch.typename(value), name))self.register_parameter(name, value)else:modules self.__dict__.get(_modules)if isinstance(value, Module):if modules is None:raise AttributeError(cannot assign module before Module.__init__() call)remove_from(self.__dict__, self._parameters, self._buffers, self._non_persistent_buffers_set)for hook in _global_module_registration_hooks.values():output hook(self, name, value)if output is not None:value outputmodules[name] valueelif modules is not None and name in modules:if value is not None:raise TypeError(cannot assign {} as child module {} (torch.nn.Module or None expected).format(torch.typename(value), name))for hook in _global_module_registration_hooks.values():output hook(self, name, value)if output is not None:value outputmodules[name] valueelse:buffers self.__dict__.get(_buffers)if buffers is not None and name in buffers:if value is not None and not isinstance(value, torch.Tensor):raise TypeError(cannot assign {} as buffer {} (torch.Tensor or None expected).format(torch.typename(value), name))for hook in _global_buffer_registration_hooks.values():output hook(self, name, value)if output is not None:value outputbuffers[name] valueelse:super().__setattr__(name, value)def __delattr__(self, name):if name in self._parameters:del self._parameters[name]elif name in self._buffers:del self._buffers[name]self._non_persistent_buffers_set.discard(name)elif name in self._modules:del self._modules[name]else:super().__delattr__(name)def _register_state_dict_hook(self, hook):rThese hooks will be called with arguments: self, state_dict,prefix, local_metadata, after the state_dict of self is set.Note that only parameters and buffers of self or its children areguaranteed to exist in state_dict. The hooks may modify state_dictinplace or return a new one.handle hooks.RemovableHandle(self._state_dict_hooks)self._state_dict_hooks[handle.id] hookreturn handledef register_state_dict_pre_hook(self, hook):rThese hooks will be called with arguments: self, prefix,and keep_vars before calling state_dict on self. The registeredhooks can be used to perform pre-processing before the state_dictcall is made.handle hooks.RemovableHandle(self._state_dict_pre_hooks)self._state_dict_pre_hooks[handle.id] hookreturn handledef _save_to_state_dict(self, destination, prefix, keep_vars):rSaves module state to destination dictionary, containing a stateof the module, but not its descendants. This is called on everysubmodule in :meth:~torch.nn.Module.state_dict.In rare cases, subclasses can achieve class-specific behavior byoverriding this method with custom logic.Args:destination (dict): a dict where state will be storedprefix (str): the prefix for parameters and buffers used in thismodulefor hook in self._state_dict_pre_hooks.values():hook(self, prefix, keep_vars)for name, param in self._parameters.items():if param is not None:destination[prefix name] param if keep_vars else param.detach()for name, buf in self._buffers.items():if buf is not None and name not in self._non_persistent_buffers_set:destination[prefix name] buf if keep_vars else buf.detach()extra_state_key prefix _EXTRA_STATE_KEY_SUFFIXif getattr(self.__class__, get_extra_state, Module.get_extra_state) is not Module.get_extra_state:destination[extra_state_key] self.get_extra_state()# The user can pass an optional arbitrary mappable object to state_dict, in which case state_dict returns# back that same object. But if they pass nothing, an OrderedDict is created and returned.T_destination TypeVar(T_destination, boundDict[str, Any])overloaddef state_dict(self, *, destination: T_destination, prefix: str ..., keep_vars: bool ...) - T_destination:...overloaddef state_dict(self, *, prefix: str ..., keep_vars: bool ...) - Dict[str, Any]:...# TODO: Change *args to * and remove the copprespinding warning in docs when BC allows.# Also remove the logic for arg parsing together.def state_dict(self, *args, destinationNone, prefix, keep_varsFalse):rReturns a dictionary containing references to the whole state of the module.Both parameters and persistent buffers (e.g. running averages) areincluded. Keys are corresponding parameter and buffer names.Parameters and buffers set to None are not included... note::The returned object is a shallow copy. It contains referencesto the modules parameters and buffers... warning::Currently state_dict() also accepts positional arguments fordestination, prefix and keep_vars in order. However,this is being deprecated and keyword arguments will be enforced infuture releases... warning::Please avoid the use of argument destination as it is notdesigned for end-users.Args:destination (dict, optional): If provided, the state of module willbe updated into the dict and the same object is returned.Otherwise, an OrderedDict will be created and returned.Default: None.prefix (str, optional): a prefix added to parameter and buffernames to compose the keys in state_dict. Default: .keep_vars (bool, optional): by default the :class:~torch.Tensor sreturned in the state dict are detached from autograd. If itsset to True, detaching will not be performed.Default: False.Returns:dict:a dictionary containing a whole state of the moduleExample:: # xdoctest: SKIP(undefined vars) module.state_dict().keys()[bias, weight]# TODO: Remove args and the parsing logic when BC allows.if len(args) 0:if destination is None:destination args[0]if len(args) 1 and prefix :prefix args[1]if len(args) 2 and keep_vars is False:keep_vars args[2]# DeprecationWarning is ignored by defaultwarnings.warn(Positional args are being deprecated, use kwargs instead. Refer to https://pytorch.org/docs/master/generated/torch.nn.Module.html#torch.nn.Module.state_dict for details.)if destination is None:destination OrderedDict()destination._metadata OrderedDict()local_metadata dict(versionself._version)if hasattr(destination, _metadata):destination._metadata[prefix[:-1]] local_metadataself._save_to_state_dict(destination, prefix, keep_vars)for name, module in self._modules.items():if module is not None:module.state_dict(destinationdestination, prefixprefix name ., keep_varskeep_vars)for hook in self._state_dict_hooks.values():hook_result hook(self, destination, prefix, local_metadata)if hook_result is not None:destination hook_resultreturn destinationdef _register_load_state_dict_pre_hook(self, hook, with_moduleFalse):rThese hooks will be called with arguments: state_dict, prefix,local_metadata, strict, missing_keys, unexpected_keys,error_msgs, before loading state_dict into self. These argumentsare exactly the same as those of _load_from_state_dict.If with_module is True, then the first argument to the hook isan instance of the module.Arguments:hook (Callable): Callable hook that will be invoked beforeloading the state dict.with_module (bool, optional): Whether or not to pass the moduleinstance to the hook as the first parameter.handle hooks.RemovableHandle(self._load_state_dict_pre_hooks)self._load_state_dict_pre_hooks[handle.id] _WrappedHook(hook, self if with_module else None)return handledef register_load_state_dict_post_hook(self, hook):rRegisters a post hook to be run after modules load_state_dictis called.It should have the following signature::hook(module, incompatible_keys) - NoneThe module argument is the current module that this hook is registeredon, and the incompatible_keys argument is a NamedTuple consistingof attributes missing_keys and unexpected_keys. missing_keysis a list of str containing the missing keys andunexpected_keys is a list of str containing the unexpected keys.The given incompatible_keys can be modified inplace if needed.Note that the checks performed when calling :func:load_state_dict withstrictTrue are affected by modifications the hook makes tomissing_keys or unexpected_keys, as expected. Additions to eitherset of keys will result in an error being thrown when strictTrue, andclearing out both missing and unexpected keys will avoid an error.Returns::class:torch.utils.hooks.RemovableHandle:a handle that can be used to remove the added hook by callinghandle.remove()handle hooks.RemovableHandle(self._load_state_dict_post_hooks)self._load_state_dict_post_hooks[handle.id] hookreturn handledef _load_from_state_dict(self, state_dict, prefix, local_metadata, strict,missing_keys, unexpected_keys, error_msgs):rCopies parameters and buffers from :attr:state_dict into onlythis module, but not its descendants. This is called on every submodulein :meth:~torch.nn.Module.load_state_dict. Metadata saved for thismodule in input :attr:state_dict is provided as :attr:local_metadata.For state dicts without metadata, :attr:local_metadata is empty.Subclasses can achieve class-specific backward compatible loading usingthe version number at local_metadata.get(version, None)... note:::attr:state_dict is not the same object as the input:attr:state_dict to :meth:~torch.nn.Module.load_state_dict. Soit can be modified.Args:state_dict (dict): a dict containing parameters andpersistent buffers.prefix (str): the prefix for parameters and buffers used in thismodulelocal_metadata (dict): a dict containing the metadata for this module.Seestrict (bool): whether to strictly enforce that the keys in:attr:state_dict with :attr:prefix match the names ofparameters and buffers in this modulemissing_keys (list of str): if strictTrue, add missing keys tothis listunexpected_keys (list of str): if strictTrue, add unexpectedkeys to this listerror_msgs (list of str): error messages should be added to thislist, and will be reported together in:meth:~torch.nn.Module.load_state_dictfor hook in self._load_state_dict_pre_hooks.values():hook(state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs)persistent_buffers {k: v for k, v in self._buffers.items() if k not in self._non_persistent_buffers_set}local_name_params itertools.chain(self._parameters.items(), persistent_buffers.items())local_state {k: v for k, v in local_name_params if v is not None}for name, param in local_state.items():key prefix nameif key in state_dict:input_param state_dict[key]if not torch.overrides.is_tensor_like(input_param):error_msgs.append(While copying the parameter named {}, expected torch.Tensor or Tensor-like object from checkpoint but received {}.format(key, type(input_param)))continue# This is used to avoid copying uninitialized parameters into# non-lazy modules, since they dont have the hook to do the checks# in such case, it will error when accessing the .shape attribute.is_param_lazy torch.nn.parameter.is_lazy(param)# Backward compatibility: loading 1-dim tensor from 0.3.* to version 0.4if not is_param_lazy and len(param.shape) 0 and len(input_param.shape) 1:input_param input_param[0]if not is_param_lazy and input_param.shape ! param.shape:# local shape should match the one in checkpointerror_msgs.append(size mismatch for {}: copying a param with shape {} from checkpoint, the shape in current model is {}..format(key, input_param.shape, param.shape))continuetry:with torch.no_grad():param.copy_(input_param)except Exception as ex:error_msgs.append(While copying the parameter named {}, whose dimensions in the model are {} and whose dimensions in the checkpoint are {}, an exception occurred : {}..format(key, param.size(), input_param.size(), ex.args))elif strict:missing_keys.append(key)extra_state_key prefix _EXTRA_STATE_KEY_SUFFIXif getattr(self.__class__, set_extra_state, Module.set_extra_state) is not Module.set_extra_state:if extra_state_key in state_dict:self.set_extra_state(state_dict[extra_state_key])elif strict:missing_keys.append(extra_state_key)elif strict and (extra_state_key in state_dict):unexpected_keys.append(extra_state_key)if strict:for key in state_dict.keys():if key.startswith(prefix) and key ! extra_state_key:input_name key[len(prefix):]input_name input_name.split(., 1)[0] # get the name of param/buffer/childif input_name not in self._modules and input_name not in local_state:unexpected_keys.append(key)def load_state_dict(self, state_dict: Mapping[str, Any],strict: bool True):rCopies parameters and buffers from :attr:state_dict intothis module and its descendants. If :attr:strict is True, thenthe keys of :attr:state_dict must exactly match the keys returnedby this modules :meth:~torch.nn.Module.state_dict function.Args:state_dict (dict): a dict containing parameters andpersistent buffers.strict (bool, optional): whether to strictly enforce that the keysin :attr:state_dict match the keys returned by this modules:meth:~torch.nn.Module.state_dict function. Default: TrueReturns:NamedTuple with missing_keys and unexpected_keys fields:* **missing_keys** is a list of str containing the missing keys* **unexpected_keys** is a list of str containing the unexpected keysNote:If a parameter or buffer is registered as None and its corresponding keyexists in :attr:state_dict, :meth:load_state_dict will raise aRuntimeError.if not isinstance(state_dict, Mapping):raise TypeError(Expected state_dict to be dict-like, got {}..format(type(state_dict)))missing_keys: List[str] []unexpected_keys: List[str] []error_msgs: List[str] []# copy state_dict so _load_from_state_dict can modify itmetadata getattr(state_dict, _metadata, None)state_dict OrderedDict(state_dict)if metadata is not None:# mypy isnt aware that _metadata exists in state_dictstate_dict._metadata metadata # type: ignore[attr-defined]def load(module, local_state_dict, prefix):local_metadata {} if metadata is None else metadata.get(prefix[:-1], {})module._load_from_state_dict(local_state_dict, prefix, local_metadata, True, missing_keys, unexpected_keys, error_msgs)for name, child in module._modules.items():if child is not None:child_prefix prefix name .child_state_dict {k: v for k, v in local_state_dict.items() if k.startswith(child_prefix)}load(child, child_state_dict, child_prefix)# Note that the hook can modify missing_keys and unexpected_keys.incompatible_keys _IncompatibleKeys(missing_keys, unexpected_keys)for hook in module._load_state_dict_post_hooks.values():out hook(module, incompatible_keys)assert out is None, (Hooks registered with register_load_state_dict_post_hook are notexpected to return new values, if incompatible_keys need to be modified,it should be done inplace.)load(self, state_dict)del loadif strict:if len(unexpected_keys) 0:error_msgs.insert(0, Unexpected key(s) in state_dict: {}. .format(, .join({}.format(k) for k in unexpected_keys)))if len(missing_keys) 0:error_msgs.insert(0, Missing key(s) in state_dict: {}. .format(, .join({}.format(k) for k in missing_keys)))if len(error_msgs) 0:raise RuntimeError(Error(s) in loading state_dict for {}:\n\t{}.format(self.__class__.__name__, \n\t.join(error_msgs)))return _IncompatibleKeys(missing_keys, unexpected_keys)def _named_members(self, get_members_fn, prefix, recurseTrue, remove_duplicate: bool True):rHelper method for yielding various names members of modules.memo set()modules self.named_modules(prefixprefix, remove_duplicateremove_duplicate) if recurse else [(prefix, self)]for module_prefix, module in modules:members get_members_fn(module)for k, v in members:if v is None or v in memo:continueif remove_duplicate:memo.add(v)name module_prefix (. if module_prefix else ) kyield name, vdef parameters(self, recurse: bool True) - Iterator[Parameter]:rReturns an iterator over module parameters.This is typically passed to an optimizer.Args:recurse (bool): if True, then yields parameters of this moduleand all submodules. Otherwise, yields only parameters thatare direct members of this module.Yields:Parameter: module parameterExample:: # xdoctest: SKIP(undefined vars) for param in model.parameters(): print(type(param), param.size())class torch.Tensor (20L,)class torch.Tensor (20L, 1L, 5L, 5L)for name, param in self.named_parameters(recurserecurse):yield paramdef named_parameters(self,prefix: str ,recurse: bool True,remove_duplicate: bool True) - Iterator[Tuple[str, Parameter]]:rReturns an iterator over module parameters, yielding both thename of the parameter as well as the parameter itself.Args:prefix (str): prefix to prepend to all parameter names.recurse (bool): if True, then yields parameters of this moduleand all submodules. Otherwise, yields only parameters thatare direct members of this module.remove_duplicate (bool, optional): whether to remove the duplicatedparameters in the result. Defaults to True.Yields:(str, Parameter): Tuple containing the name and parameterExample:: # xdoctest: SKIP(undefined vars) for name, param in self.named_parameters(): if name in [bias]: print(param.size())gen self._named_members(lambda module: module._parameters.items(),prefixprefix, recurserecurse, remove_duplicateremove_duplicate)yield from gendef buffers(self, recurse: bool True) - Iterator[Tensor]:rReturns an iterator over module buffers.Args:recurse (bool): if True, then yields buffers of this moduleand all submodules. Otherwise, yields only buffers thatare direct members of this module.Yields:torch.Tensor: module bufferExample:: # xdoctest: SKIP(undefined vars) for buf in model.buffers(): print(type(buf), buf.size())class torch.Tensor (20L,)class torch.Tensor (20L, 1L, 5L, 5L)for _, buf in self.named_buffers(recurserecurse):yield bufdef named_buffers(self, prefix: str , recurse: bool True, remove_duplicate: bool True) - Iterator[Tuple[str, Tensor]]:rReturns an iterator over module buffers, yielding both thename of the buffer as well as the buffer itself.Args:prefix (str): prefix to prepend to all buffer names.recurse (bool, optional): if True, then yields buffers of this moduleand all submodules. Otherwise, yields only buffers thatare direct members of this module. Defaults to True.remove_duplicate (bool, optional): whether to remove the duplicated buffers in the result. Defaults to True.Yields:(str, torch.Tensor): Tuple containing the name and bufferExample:: # xdoctest: SKIP(undefined vars) for name, buf in self.named_buffers(): if name in [running_var]: print(buf.size())gen self._named_members(lambda module: module._buffers.items(),prefixprefix, recurserecurse, remove_duplicateremove_duplicate)yield from gendef children(self) - Iterator[Module]:rReturns an iterator over immediate children modules.Yields:Module: a child modulefor name, module in self.named_children():yield moduledef named_children(self) - Iterator[Tuple[str, Module]]:rReturns an iterator over immediate children modules, yielding boththe name of the module as well as the module itself.Yields:(str, Module): Tuple containing a name and child moduleExample:: # xdoctest: SKIP(undefined vars) for name, module in model.named_children(): if name in [conv4, conv5]: print(module)memo set()for name, module in self._modules.items():if module is not None and module not in memo:memo.add(module)yield name, moduledef modules(self) - Iterator[Module]:rReturns an iterator over all modules in the network.Yields:Module: a module in the networkNote:Duplicate modules are returned only once. In the followingexample, l will be returned only once.Example:: l nn.Linear(2, 2) net nn.Sequential(l, l) for idx, m in enumerate(net.modules()):... print(idx, -, m)0 - Sequential((0): Linear(in_features2, out_features2, biasTrue)(1): Linear(in_features2, out_features2, biasTrue))1 - Linear(in_features2, out_features2, biasTrue)for _, module in self.named_modules():yield moduledef named_modules(self, memo: Optional[Set[Module]] None, prefix: str , remove_duplicate: bool True):rReturns an iterator over all modules in the network, yieldingboth the name of the module as well as the module itself.Args:memo: a memo to store the set of modules already added to the resultprefix: a prefix that will be added to the name of the moduleremove_duplicate: whether to remove the duplicated module instances in the resultor notYields:(str, Module): Tuple of name and moduleNote:Duplicate modules are returned only once. In the followingexample, l will be returned only once.Example:: l nn.Linear(2, 2) net nn.Sequential(l, l) for idx, m in enumerate(net.named_modules()):... print(idx, -, m)0 - (, Sequential((0): Linear(in_features2, out_features2, biasTrue)(1): Linear(in_features2, out_features2, biasTrue)))1 - (0, Linear(in_features2, out_features2, biasTrue))if memo is None:memo set()if self not in memo:if remove_duplicate:memo.add(self)yield prefix, selffor name, module in self._modules.items():if module is None:continuesubmodule_prefix prefix (. if prefix else ) namefor m in module.named_modules(memo, submodule_prefix, remove_duplicate):yield mdef train(self: T, mode: bool True) - T:rSets the module in training mode.This has any effect only on certain modules. See documentations ofparticular modules for details of their behaviors in training/evaluationmode, if they are affected, e.g. :class:Dropout, :class:BatchNorm,etc.Args:mode (bool): whether to set training mode (True) or evaluationmode (False). Default: True.Returns:Module: selfif not isinstance(mode, bool):raise ValueError(training mode is expected to be boolean)self.training modefor module in self.children():module.train(mode)return selfdef eval(self: T) - T:rSets the module in evaluation mode.This has any effect only on certain modules. See documentations ofparticular modules for details of their behaviors in training/evaluationmode, if they are affected, e.g. :class:Dropout, :class:BatchNorm,etc.This is equivalent with :meth:self.train(False) torch.nn.Module.train.See :ref:locally-disable-grad-doc for a comparison between.eval() and several similar mechanisms that may be confused with it.Returns:Module: selfreturn self.train(False)def requires_grad_(self: T, requires_grad: bool True) - T:rChange if autograd should record operations on parameters in thismodule.This method sets the parameters :attr:requires_grad attributesin-place.This method is helpful for freezing part of the module for finetuningor training parts of a model individually (e.g., GAN training).See :ref:locally-disable-grad-doc for a comparison between.requires_grad_() and several similar mechanisms that may be confused with it.Args:requires_grad (bool): whether autograd should record operations onparameters in this module. Default: True.Returns:Module: selffor p in self.parameters():p.requires_grad_(requires_grad)return selfdef zero_grad(self, set_to_none: bool True) - None:rSets gradients of all model parameters to zero. See similar functionunder :class:torch.optim.Optimizer for more context.Args:set_to_none (bool): instead of setting to zero, set the grads to None.See :meth:torch.optim.Optimizer.zero_grad for details.if getattr(self, _is_replica, False):warnings.warn(Calling .zero_grad() from a module created with nn.DataParallel() has no effect. The parameters are copied (in a differentiable manner) from the original module. This means they are not leaf nodes in autograd and so dont accumulate gradients. If you need gradients in your forward method, consider using autograd.grad instead.)for p in self.parameters():if p.grad is not None:if set_to_none:p.grad Noneelse:if p.grad.grad_fn is not None:p.grad.detach_()else:p.grad.requires_grad_(False)p.grad.zero_()def share_memory(self: T) - T:rSee :meth:torch.Tensor.share_memory_return self._apply(lambda t: t.share_memory_())def _get_name(self):return self.__class__.__name__def extra_repr(self) - str:rSet the extra representation of the moduleTo print customized extra information, you should re-implementthis method in your own modules. Both single-line and multi-linestrings are acceptable.return def __repr__(self):# We treat the extra repr like the sub-module, one item per lineextra_lines []extra_repr self.extra_repr()# empty string will be split into list []if extra_repr:extra_lines extra_repr.split(\n)child_lines []for key, module in self._modules.items():mod_str repr(module)mod_str _addindent(mod_str, 2)child_lines.append(( key ): mod_str)lines extra_lines child_linesmain_str self._get_name() (if lines:# simple one-liner info, which most builtin Modules will useif len(extra_lines) 1 and not child_lines:main_str extra_lines[0]else:main_str \n \n .join(lines) \nmain_str )return main_strdef __dir__(self):module_attrs dir(self.__class__)attrs list(self.__dict__.keys())parameters list(self._parameters.keys())modules list(self._modules.keys())buffers list(self._buffers.keys())keys module_attrs attrs parameters modules buffers# Eliminate attrs that are not legal Python variable nameskeys [key for key in keys if not key[0].isdigit()]return sorted(keys)def _replicate_for_data_parallel(self):replica self.__new__(type(self))replica.__dict__ self.__dict__.copy()# replicas do not have parameters themselves, the replicas reference the original# module.replica._parameters OrderedDict()replica._buffers replica._buffers.copy()replica._modules replica._modules.copy()replica._is_replica True # type: ignore[assignment]return replica
