StochasticMuZeroModel

class lzero.model.stochastic_muzero_model.StochasticMuZeroModel(observation_shape: SequenceType = (12, 96, 96), action_space_size: int = 6, chance_space_size: int = 2, num_res_blocks: int = 1, num_channels: int = 64, reward_head_channels: int = 16, value_head_channels: int = 16, policy_head_channels: int = 16, fc_reward_layers: SequenceType = [32], fc_value_layers: SequenceType = [32], fc_policy_layers: SequenceType = [32], reward_support_size: int = 601, value_support_size: int = 601, proj_hid: int = 1024, proj_out: int = 1024, pred_hid: int = 512, pred_out: int = 1024, self_supervised_learning_loss: bool = False, categorical_distribution: bool = True, activation: Module = ReLU(inplace=True), last_linear_layer_init_zero: bool = True, state_norm: bool = False, downsample: bool = False, *args, **kwargs)

Bases: Module

T_destination = ~T_destination

__init__(observation_shape: SequenceType = (12, 96, 96), action_space_size: int = 6, chance_space_size: int = 2, num_res_blocks: int = 1, num_channels: int = 64, reward_head_channels: int = 16, value_head_channels: int = 16, policy_head_channels: int = 16, fc_reward_layers: SequenceType = [32], fc_value_layers: SequenceType = [32], fc_policy_layers: SequenceType = [32], reward_support_size: int = 601, value_support_size: int = 601, proj_hid: int = 1024, proj_out: int = 1024, pred_hid: int = 512, pred_out: int = 1024, self_supervised_learning_loss: bool = False, categorical_distribution: bool = True, activation: Module = ReLU(inplace=True), last_linear_layer_init_zero: bool = True, state_norm: bool = False, downsample: bool = False, *args, **kwargs)

Overview:

The definition of the neural network model used in Stochastic MuZero, which is proposed in the paper https://openreview.net/pdf?id=X6D9bAHhBQ1. Stochastic MuZero model consists of a representation network, a dynamics network and a prediction network. The networks are built on convolution residual blocks and fully connected layers.

Parameters:

• observation_shape (-) – Observation space shape, e.g. [C, W, H]=[12, 96, 96] for Atari.

• action_space_size (-) – (int): Action space size, usually an integer number for discrete action space.

• chance_space_size (-) – (int): Chance space size, the action space for decision node, usually an integer number for discrete action space.

• num_res_blocks (-) – The number of res blocks in AlphaZero model.

• num_channels (-) – The channels of hidden states.

• reward_head_channels (-) – The channels of reward head.

• value_head_channels (-) – The channels of value head.

• policy_head_channels (-) – The channels of policy head.

• fc_reward_layers (-) – The number of hidden layers of the reward head (MLP head).

• fc_value_layers (-) – The number of hidden layers used in value head (MLP head).

• fc_policy_layers (-) – The number of hidden layers used in policy head (MLP head).

• reward_support_size (-) – The size of categorical reward output

• value_support_size (-) – The size of categorical value output.

• proj_hid (-) – The size of projection hidden layer.

• proj_out (-) – The size of projection output layer.

• pred_hid (-) – The size of prediction hidden layer.

• pred_out (-) – The size of prediction output layer.

• self_supervised_learning_loss (-) – Whether to use self_supervised_learning related networks in Stochastic MuZero model, default set it to False.

• categorical_distribution (-) – Whether to use discrete support to represent categorical distribution for value and reward.

• activation (-) – Activation function used in network, which often use in-place operation to speedup, e.g. ReLU(inplace=True).

• last_linear_layer_init_zero (-) – Whether to use zero initialization for the last layer of dynamics/prediction mlp, default set it to True.

• state_norm (-) – Whether to use normalization for hidden states, default set it to False.

• downsample (-) – Whether to do downsampling for observations in representation_network, defaults to True. This option is often used in video games like Atari. In board games like go, we don’t need this module.

_afterstate_dynamics(latent_state: Tensor, action: Tensor) → Tuple[Tensor, Tensor]

Overview:

Concatenate latent_state and action and use the dynamics network to predict next_latent_state and reward.

Parameters:

• latent_state (-) – The encoding latent state of input state.

• action (-) – The predicted action to rollout.

Returns:

The predicted latent state of the next timestep. - reward (torch.Tensor): The predicted reward of the current latent state and selected action.

Return type:

• next_latent_state (torch.Tensor)

Shapes:

• latent_state (torch.Tensor): \((B, H_, W_)\), where B is batch_size, H_ is the height of latent state, W_ is the width of latent state.

• action (torch.Tensor): \((B, )\), where B is batch_size.

• next_latent_state (torch.Tensor): \((B, H_, W_)\), where B is batch_size, H_ is the height of latent state, W_ is the width of latent state.

• reward (torch.Tensor): \((B, reward_support_size)\), where B is batch_size.

_afterstate_prediction(afterstate: Tensor) → Tuple[Tensor, Tensor]

Overview:

Use the prediction network to predict policy_logits and value.

Parameters:

latent_state (-) – The encoding latent state of input state.

Returns:

The output logit to select discrete action. - value (torch.Tensor): The output value of input state to help policy improvement and evaluation.

Return type:

• policy_logits (torch.Tensor)

Shapes:

• latent_state (torch.Tensor): \((B, H_, W_)\), where B is batch_size, H_ is the height of latent state, W_ is the width of latent state.

• policy_logits (torch.Tensor): \((B, action_dim)\), where B is batch_size.

• value (torch.Tensor): \((B, value_support_size)\), where B is batch_size.

_apply(fn, recurse=True)

_backward_hooks: Dict[int, Callable]

_backward_pre_hooks: Dict[int, Callable]

_buffers: Dict[str, Tensor | None]

_call_impl(*args, **kwargs)

_compiled_call_impl: Callable | None = None

_dynamics(latent_state: Tensor, action: Tensor) → Tuple[Tensor, Tensor]

Overview:

Concatenate latent_state and action and use the dynamics network to predict next_latent_state and reward.

Parameters:

• latent_state (-) – The encoding latent state of input state.

• action (-) – The predicted action to rollout.

Returns:

The predicted latent state of the next timestep. - reward (torch.Tensor): The predicted reward of the current latent state and selected action.

Return type:

• next_latent_state (torch.Tensor)

Shapes:

• latent_state (torch.Tensor): \((B, H_, W_)\), where B is batch_size, H_ is the height of latent state, W_ is the width of latent state.

• action (torch.Tensor): \((B, )\), where B is batch_size.

• next_latent_state (torch.Tensor): \((B, H_, W_)\), where B is batch_size, H_ is the height of latent state, W_ is the width of latent state.

• reward (torch.Tensor): \((B, reward_support_size)\), where B is batch_size.

_forward_hooks: Dict[int, Callable]

_forward_hooks_always_called: Dict[int, bool]

_forward_hooks_with_kwargs: Dict[int, bool]

_forward_pre_hooks: Dict[int, Callable]

_forward_pre_hooks_with_kwargs: Dict[int, bool]

_get_backward_hooks()

Return the backward hooks for use in the call function.

It returns two lists, one with the full backward hooks and one with the non-full backward hooks.

_get_backward_pre_hooks()

_get_name()

_is_full_backward_hook: bool | None

_load_from_state_dict(state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs)

Copy parameters and buffers from state_dict into only this module, but not its descendants.

This is called on every submodule in load_state_dict(). Metadata saved for this module in input state_dict is provided as local_metadata. For state dicts without metadata, local_metadata is empty. Subclasses can achieve class-specific backward compatible loading using the version number at local_metadata.get(“version”, None). Additionally, local_metadata can also contain the key assign_to_params_buffers that indicates whether keys should be assigned their corresponding tensor in the state_dict.

Note

state_dict is not the same object as the input state_dict to load_state_dict(). So it can be modified.

Parameters:

• state_dict (dict) – a dict containing parameters and persistent buffers.

• prefix (str) – the prefix for parameters and buffers used in this module

• local_metadata (dict) – a dict containing the metadata for this module. See

• strict (bool) – whether to strictly enforce that the keys in state_dict with prefix match the names of parameters and buffers in this module

• missing_keys (list of str) – if strict=True, add missing keys to this list

• unexpected_keys (list of str) – if strict=True, add unexpected keys to this list

• error_msgs (list of str) – error messages should be added to this list, and will be reported together in load_state_dict()

_load_state_dict_post_hooks: Dict[int, Callable]

_load_state_dict_pre_hooks: Dict[int, Callable]

_maybe_warn_non_full_backward_hook(inputs, result, grad_fn)

_modules: Dict[str, 'Module' | None]

_named_members(get_members_fn, prefix='', recurse=True, remove_duplicate: bool = True)

Help yield various names + members of modules.

_non_persistent_buffers_set: Set[str]

_parameters: Dict[str, Parameter | None]

_prediction(latent_state: Tensor) → Tuple[Tensor, Tensor]

Overview:

Use the prediction network to predict policy_logits and value.

Parameters:

latent_state (-) – The encoding latent state of input state.

Returns:

The output logit to select discrete action. - value (torch.Tensor): The output value of input state to help policy improvement and evaluation.

Return type:

• policy_logits (torch.Tensor)

Shapes:

• latent_state (torch.Tensor): \((B, H_, W_)\), where B is batch_size, H_ is the height of latent state, W_ is the width of latent state.

• policy_logits (torch.Tensor): \((B, action_dim)\), where B is batch_size.

• value (torch.Tensor): \((B, value_support_size)\), where B is batch_size.

_register_load_state_dict_pre_hook(hook, with_module=False)

See register_load_state_dict_pre_hook() for details.

A subtle difference is that if with_module is set to False, then the hook will not take the module as the first argument whereas register_load_state_dict_pre_hook() always takes the module as the first argument.

Parameters:

• hook (Callable) – Callable hook that will be invoked before loading the state dict.

• with_module (bool, optional) – Whether or not to pass the module instance to the hook as the first parameter.

_register_state_dict_hook(hook)

Register a post-hook for the state_dict() method.

It should have the following signature::

hook(module, state_dict, prefix, local_metadata) -> None or state_dict

The registered hooks can modify the state_dict inplace or return a new one. If a new state_dict is returned, it will only be respected if it is the root module that state_dict() is called from.

_replicate_for_data_parallel()

_representation(observation: Tensor) → Tensor

Overview:

Use the representation network to encode the observations into latent state.

Parameters:

obs (-) – The 2D image observation data.

Returns:

The encoding latent state of input state.

Return type:

• latent_state (torch.Tensor)

Shapes:

• obs (torch.Tensor): \((B, num_channel, obs_shape[1], obs_shape[2])\), where B is batch_size.

• latent_state (torch.Tensor): \((B, H_, W_)\), where B is batch_size, H_ is the height of latent state, W_ is the width of latent state.

_save_to_state_dict(destination, prefix, keep_vars)

Save module state to the destination dictionary.

The destination dictionary will contain the state of the module, but not its descendants. This is called on every submodule in state_dict().

In rare cases, subclasses can achieve class-specific behavior by overriding this method with custom logic.

Parameters:

• destination (dict) – a dict where state will be stored

• prefix (str) – the prefix for parameters and buffers used in this module

_slow_forward(*input, **kwargs)

_state_dict_hooks: Dict[int, Callable]

_state_dict_pre_hooks: Dict[int, Callable]

_version: int = 1

This allows better BC support for load_state_dict(). In state_dict(), the version number will be saved as in the attribute _metadata of the returned state dict, and thus pickled. _metadata is a dictionary 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 shall be bumped, and the module’s _load_from_state_dict method can compare the version number and do appropriate changes if the state dict is from before the change.

_wrapped_call_impl(*args, **kwargs)

add_module(name: str, module: Module | None) → None

Add a child module to the current module.

The module can be accessed as an attribute using the given name.

Parameters:

• name (str) – name of the child module. The child module can be accessed from this module using the given name

• module (Module) – child module to be added to the module.

apply(fn: Callable[[Module], None]) → T

Apply fn recursively to every submodule (as returned by .children()) as well as self.

Typical use includes initializing the parameters of a model (see also nn-init-doc).

Parameters:

fn (Module -> None) – function to be applied to each submodule

Returns:

self

Return type:

Module

Example:

>>> @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_features=2, out_features=2, bias=True)
Parameter containing:
tensor([[1., 1.],
        [1., 1.]], requires_grad=True)
Linear(in_features=2, out_features=2, bias=True)
Parameter containing:
tensor([[1., 1.],
        [1., 1.]], requires_grad=True)
Sequential(
  (0): Linear(in_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
)

bfloat16() → T

Casts all floating point parameters and buffers to bfloat16 datatype.

Note

This method modifies the module in-place.

Returns:

self

Return type:

Module

buffers(recurse: bool = True) → Iterator[Tensor]

Return an iterator over module buffers.

Parameters:

recurse (bool) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module.

Yields:

torch.Tensor – module buffer

Example:

>>> # 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)

call_super_init: bool = False

chance_encode(observation: Tensor)

children() → Iterator[Module]

Return 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

Move all model parameters and buffers to the CPU.

Note

This method modifies the module in-place.

Returns:

self

Return type:

Module

cuda(device: int | device | None = None) → T

Move 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.

Parameters:

device (int, optional) – if specified, all parameters will be copied to that device

Returns:

self

Return type:

Module

double() → T

Casts all floating point parameters and buffers to double datatype.

Note

This method modifies the module in-place.

Returns:

self

Return type:

Module

dump_patches: bool = False

eval() → T

Set 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-disable-grad-doc for a comparison between .eval() and several similar mechanisms that may be confused with it.

Returns:

self

Return type:

Module

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:

Module

forward(*input: Any) → None

Define the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

get_buffer(target: str) → Tensor

Return the buffer given by target if it exists, otherwise throw an error.

See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target.

Parameters:

target – The fully-qualified string name of the buffer to look for. (See get_submodule for how to specify a fully-qualified string.)

Returns:

The buffer referenced by target

Return type:

torch.Tensor

Raises:

AttributeError – If the target string references an invalid path or resolves to something that is not a buffer

get_extra_state() → Any

Return 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:

object

get_parameter(target: str) → Parameter

Return the parameter given by target if it exists, otherwise throw an error.

See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target.

Parameters:

target – The fully-qualified string name of the Parameter to look for. (See get_submodule for how to specify a fully-qualified string.)

Returns:

The Parameter referenced by target

Return type:

torch.nn.Parameter

Raises:

AttributeError – If the target string references an invalid path or resolves to something that is not an nn.Parameter

get_params_mean() → float

get_submodule(target: str) → Module

Return the submodule given by target if it exists, otherwise throw 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 submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.)

To check whether or not we have the linear submodule, we would call get_submodule("net_b.linear"). To check whether we have the conv submodule, we would call get_submodule("net_b.net_c.conv").

The runtime of get_submodule is bounded by the degree of module nesting in target. A query against named_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:

torch.nn.Module

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:

Module

initial_inference(obs: Tensor) → MZNetworkOutput

Overview:

Initial inference of Stochastic MuZero model, which is the first step of the Stochastic MuZero model. To perform the initial inference, we first use the representation network to obtain the latent_state. Then we use the prediction network to predict value and policy_logits of the latent_state.

Parameters:

obs (-) – The 2D image observation data.

Returns (MZNetworkOutput):

• value (torch.Tensor): The output value of input state to help policy improvement and evaluation.

• reward (torch.Tensor): The predicted reward of input state and selected action. In initial inference, we set it to zero vector.

• policy_logits (torch.Tensor): The output logit to select discrete action.

• latent_state (torch.Tensor): The encoding latent state of input state.

Shapes:

• obs (torch.Tensor): \((B, num_channel, obs_shape[1], obs_shape[2])\), where B is batch_size.

• value (torch.Tensor): \((B, value_support_size)\), where B is batch_size.

• reward (torch.Tensor): \((B, reward_support_size)\), where B is batch_size.

• policy_logits (torch.Tensor): \((B, action_dim)\), where B is batch_size.

• latent_state (torch.Tensor): \((B, H_, W_)\), where B is batch_size, H_ is the height of latent state, W_ is the width of latent state.

ipu(device: int | device | None = None) → T

Move 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.

Parameters:

device (int, optional) – if specified, all parameters will be copied to that device

Returns:

self

Return type:

Module

load_state_dict(state_dict: Mapping[str, Any], strict: bool = True, assign: bool = False)

Copy parameters and buffers from state_dict into this module and its descendants.

If strict is True, then the keys of state_dict must exactly match the keys returned by this module’s state_dict() function.

Warning

If assign is True the optimizer must be created after the call to load_state_dict unless get_swap_module_params_on_conversion() is True.

Parameters:

• state_dict (dict) – a dict containing parameters and persistent buffers.

• strict (bool, optional) – whether to strictly enforce that the keys in state_dict match the keys returned by this module’s state_dict() function. Default: True

• assign (bool, optional) – When False, the properties of the tensors in the current module are preserved while when True, the properties of the Tensors in the state dict are preserved. The only exception is the requires_grad field of Default: ``False`

Returns:

• missing_keys is a list of str containing any keys that are expected

  by this module but missing from the provided state_dict.

• unexpected_keys is a list of str containing the keys that are not

  expected by this module but present in the provided state_dict.

Return type:

NamedTuple with missing_keys and unexpected_keys fields

Note

If a parameter or buffer is registered as None and its corresponding key exists in state_dict, load_state_dict() will raise a RuntimeError.

modules() → Iterator[Module]

Return an iterator over all modules in the network.

Yields:

Module – a module in the network

Note

Duplicate modules are returned only once. In the following example, 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_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
)
1 -> Linear(in_features=2, out_features=2, bias=True)

mtia(device: int | device | None = None) → T

Move all model parameters and buffers to the MTIA.

This also makes associated parameters and buffers different objects. So it should be called before constructing optimizer if the module will live on MTIA while being optimized.

Note

This method modifies the module in-place.

Parameters:

device (int, optional) – if specified, all parameters will be copied to that device

Returns:

self

Return type:

Module

named_buffers(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[Tuple[str, Tensor]]

Return an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.

Parameters:

• prefix (str) – prefix to prepend to all buffer names.

• recurse (bool, optional) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are 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 buffer

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, buf in self.named_buffers():
>>>     if name in ['running_var']:
>>>         print(buf.size())

named_children() → Iterator[Tuple[str, Module]]

Return an iterator over immediate children modules, yielding both the name of the module as well as the module itself.

Yields:

(str, Module) – Tuple containing a name and child module

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, module in model.named_children():
>>>     if name in ['conv4', 'conv5']:
>>>         print(module)

named_modules(memo: Set[Module] | None = None, prefix: str = '', remove_duplicate: bool = True)

Return an iterator over all modules in the network, yielding both the name of the module as well as the module itself.

Parameters:

• memo – a memo to store the set of modules already added to the result

• prefix – a prefix that will be added to the name of the module

• remove_duplicate – whether to remove the duplicated module instances in the result or not

Yields:

(str, Module) – Tuple of name and module

Note

Duplicate modules are returned only once. In the following example, 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_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
))
1 -> ('0', Linear(in_features=2, out_features=2, bias=True))

named_parameters(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[Tuple[str, Parameter]]

Return an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.

Parameters:

• prefix (str) – prefix to prepend to all parameter names.

• recurse (bool) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module.

• remove_duplicate (bool, optional) – whether to remove the duplicated parameters in the result. Defaults to True.

Yields:

(str, Parameter) – Tuple containing the name and parameter

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, param in self.named_parameters():
>>>     if name in ['bias']:
>>>         print(param.size())

parameters(recurse: bool = True) → Iterator[Parameter]

Return an iterator over module parameters.

This is typically passed to an optimizer.

Parameters:

recurse (bool) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module.

Yields:

Parameter – module parameter

Example:

>>> # 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)

project(latent_state: Tensor, with_grad: bool = True) → Tensor

Overview:

Project the latent state to a lower dimension to calculate the self-supervised loss, which is involved in in EfficientZero. For more details, please refer to paper Exploring Simple Siamese Representation Learning.

Parameters:

• latent_state (-) – The encoding latent state of input state.

• with_grad (-) – Whether to calculate gradient for the projection result.

Returns:

The result embedding vector of projection operation.

Return type:

• proj (torch.Tensor)

Shapes:

• latent_state (torch.Tensor): \((B, H_, W_)\), where B is batch_size, H_ is the height of latent state, W_ is the width of latent state.

• proj (torch.Tensor): \((B, projection_output_dim)\), where B is batch_size.

Examples

>>> latent_state = torch.randn(256, 64, 6, 6)
>>> output = self.project(latent_state)
>>> output.shape # (256, 1024)

Note

for Atari: observation_shape = (12, 96, 96), # original shape is (3,96,96), frame_stack_num=4 if downsample is True, latent_state.shape: (batch_size, num_channel, obs_shape[1] / 16, obs_shape[2] / 16) i.e., (256, 64, 96 / 16, 96 / 16) = (256, 64, 6, 6) latent_state reshape: (256, 64, 6, 6) -> (256,64*6*6) = (256, 2304) # self.projection_input_dim = 64*6*6 = 2304 # self.projection_output_dim = 1024

recurrent_inference(state: Tensor, option: Tensor, afterstate: bool = False) → MZNetworkOutput

Overview:

Recurrent inference of Stochastic MuZero model, which is the rollout step of the Stochastic MuZero model. To perform the recurrent inference, we first use the dynamics network to predict next_latent_state, reward, by the given current latent_state and action. We then use the prediction network to predict the value and policy_logits of the current latent_state.

Parameters:

• state (-) – The encoding latent state of input state or the afterstate.

• option (-) – The action to rollout or the chance to predict next latent state.

• afterstate (-) – Whether to use afterstate prediction network to predict next latent state.

Returns (MZNetworkOutput):

• value (torch.Tensor): The output value of input state to help policy improvement and evaluation.

• reward (torch.Tensor): The predicted reward of input state and selected action.

• policy_logits (torch.Tensor): The output logit to select discrete action.

• latent_state (torch.Tensor): The encoding latent state of input state.

• next_latent_state (torch.Tensor): The predicted next latent state.

Shapes:

• obs (torch.Tensor): \((B, num_channel, obs_shape[1], obs_shape[2])\), where B is batch_size.

• action (torch.Tensor): \((B, )\), where B is batch_size.

• value (torch.Tensor): \((B, value_support_size)\), where B is batch_size.

• reward (torch.Tensor): \((B, reward_support_size)\), where B is batch_size.

• policy_logits (torch.Tensor): \((B, action_dim)\), where B is batch_size.

• latent_state (torch.Tensor): \((B, H_, W_)\), where B is batch_size, H_ is the height of latent state, W_ is the width of latent state.

• next_latent_state (torch.Tensor): \((B, H_, W_)\), where B is batch_size, H_ is the height of latent state, W_ is the width of latent state.

register_backward_hook(hook: Callable[[Module, Tuple[Tensor, ...] | Tensor, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor]) → RemovableHandle

Register 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_buffer(name: str, tensor: Tensor | None, persistent: bool = True) → None

Add a buffer to the module.

This is typically used to register a buffer that should not to be considered a model parameter. For example, BatchNorm’s running_mean is not a parameter, but is part of the module’s state. Buffers, by default, are persistent and will be saved alongside parameters. This behavior can be changed by setting persistent to False. The only difference between a persistent buffer and a non-persistent buffer is that the latter will not be a part of this module’s state_dict.

Buffers can be accessed as attributes using given names.

Parameters:

• name (str) – name of the buffer. The buffer can be accessed from this module using the given name

• tensor (Tensor or None) – buffer to be registered. If None, then operations that run on buffers, such as cuda, are ignored. If None, the buffer is not included in the module’s state_dict.

• persistent (bool) – whether the buffer is part of this module’s state_dict.

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> self.register_buffer('running_mean', torch.zeros(num_features))

register_forward_hook(hook: Callable[[T, Tuple[Any, ...], Any], Any | None] | Callable[[T, Tuple[Any, ...], Dict[str, Any], Any], Any | None], *, prepend: bool = False, with_kwargs: bool = False, always_call: bool = False) → RemovableHandle

Register a forward hook on the module.

The hook will be called every time after forward() has computed an output.

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 the forward. The hook can modify the output. It can modify the input inplace but it will not have effect on forward since this is called after forward() is called. The hook should have the following signature:

hook(module, args, output) -> None or modified output

If with_kwargs is True, the forward hook will be passed the kwargs 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 provided hook will be fired before all existing forward hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing forward hooks on this torch.nn.modules.Module. Note that global forward hooks registered with register_module_forward_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

• always_call (bool) – If True the hook 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: Callable[[T, Tuple[Any, ...]], Any | None] | Callable[[T, Tuple[Any, ...], Dict[str, Any]], Tuple[Any, Dict[str, Any]] | None], *, prepend: bool = False, with_kwargs: bool = False) → RemovableHandle

Register 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 the forward. 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 existing forward_pre hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing forward_pre hooks on this torch.nn.modules.Module. Note that global forward_pre hooks registered with register_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, Tuple[Tensor, ...] | Tensor, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle

Register 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 and grad_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 of grad_input in subsequent computations. grad_input will only correspond to the inputs given as positional arguments and all kwarg arguments are ignored. Entries in grad_input and grad_output will be None 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 existing backward hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing backward hooks on this torch.nn.modules.Module. Note that global backward hooks registered with register_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, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle

Register 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 of grad_output in subsequent computations. Entries in grad_output will be None 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 existing backward_pre hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing backward_pre hooks on this torch.nn.modules.Module. Note that global backward_pre hooks registered with register_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)

Register 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 the incompatible_keys argument is a NamedTuple consisting of attributes missing_keys and unexpected_keys. missing_keys is a list of str containing the missing keys and unexpected_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 load_state_dict() with strict=True are affected by modifications the hook makes to missing_keys or unexpected_keys, as expected. Additions to either set of keys will result in an error being thrown when strict=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_load_state_dict_pre_hook(hook)

Register a pre-hook to be run before module’s load_state_dict() is called.

It should have the following signature::

hook(module, state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs) -> None # noqa: B950

Parameters:

hook (Callable) – Callable hook that will be invoked before loading the state dict.

register_module(name: str, module: Module | None) → None

Alias for add_module().

register_parameter(name: str, param: Parameter | None) → None

Add a parameter to the module.

The parameter can be accessed as an attribute using given name.

Parameters:

• name (str) – name of the parameter. The parameter can be accessed from this module using the given name

• param (Parameter or None) – parameter to be added to the module. If None, then operations that run on parameters, such as cuda, are ignored. If None, the parameter is not included in the module’s state_dict.

register_state_dict_post_hook(hook)

Register a post-hook for the state_dict() method.

It should have the following signature::

hook(module, state_dict, prefix, local_metadata) -> None

The registered hooks can modify the state_dict inplace.

register_state_dict_pre_hook(hook)

Register a pre-hook for the state_dict() method.

It should have the following signature::

hook(module, prefix, keep_vars) -> None

The registered hooks can be used to perform pre-processing before the state_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-disable-grad-doc for a comparison between .requires_grad_() and several similar mechanisms that may be confused with it.

Parameters:

requires_grad (bool) – whether autograd should record operations on parameters in this module. Default: True.

Returns:

self

Return type:

Module

set_extra_state(state: Any) → None

Set extra state contained in the loaded state_dict.

This function is called from load_state_dict() to handle any extra state found within the state_dict. Implement this function and a corresponding get_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

set_submodule(target: str, module: Module) → None

Set the submodule given by target if it exists, otherwise throw 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 submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.)

To overide the Conv2d with a new submodule Linear, you would call set_submodule("net_b.net_c.conv", nn.Linear(33, 16)).

Parameters:

• target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.)

• module – The module to set the submodule to.

Raises:

• ValueError – If the target string is empty

• AttributeError – If the target string references an invalid path or resolves to something that is not an nn.Module

share_memory() → T

See torch.Tensor.share_memory_().

state_dict(*args, destination=None, prefix='', keep_vars=False)

Return a dictionary containing references to the whole state of the module.

Both parameters and persistent buffers (e.g. running averages) are included. 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 references to the module’s parameters and buffers.

Warning

Currently state_dict() also accepts positional arguments for destination, prefix and keep_vars in order. However, this is being deprecated and keyword arguments will be enforced in future releases.

Warning

Please avoid the use of argument destination as it is not designed for end-users.

Parameters:

• destination (dict, optional) – If provided, the state of module will be 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 buffer names to compose the keys in state_dict. Default: ''.

• keep_vars (bool, optional) – by default the Tensor s returned in the state dict are detached from autograd. If it’s set to True, detaching will not be performed. Default: False.

Returns:

a dictionary containing a whole state of the module

Return type:

dict

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> module.state_dict().keys()
['bias', 'weight']

to(*args, **kwargs)

Move and/or cast 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 complex dtypes. In addition, this method will only cast the floating point or complex parameters and buffers to dtype (if given). The integral parameters and buffers will be moved device, if that is given, but with dtypes unchanged. When non_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 module

• dtype (torch.dtype) – the desired floating point or complex dtype of the parameters and buffers in this module

• tensor (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:

Module

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: int | str | device | None, recurse: bool = True) → T

Move 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:

Module

train(mode: bool = True) → T

Set the module in training 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.

Parameters:

mode (bool) – whether to set training mode (True) or evaluation mode (False). Default: True.

Returns:

self

Return type:

Module

training: bool

type(dst_type: dtype | str) → T

Casts all parameters and buffers to dst_type.

Note

This method modifies the module in-place.

Parameters:

dst_type (type or string) – the desired type

Returns:

self

Return type:

Module

xpu(device: int | device | None = None) → T

Move 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.

Parameters:

device (int, optional) – if specified, all parameters will be copied to that device

Returns:

self

Return type:

Module

zero_grad(set_to_none: bool = True) → None

Reset 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.

class lzero.model.stochastic_muzero_model.DynamicsNetwork(num_res_blocks: int, num_channels: int, reward_head_channels: int, fc_reward_layers: SequenceType, output_support_size: int, flatten_output_size_for_reward_head: int, last_linear_layer_init_zero: bool = True, activation: Module | None = ReLU(inplace=True))

Bases: Module

T_destination = ~T_destination

__init__(num_res_blocks: int, num_channels: int, reward_head_channels: int, fc_reward_layers: SequenceType, output_support_size: int, flatten_output_size_for_reward_head: int, last_linear_layer_init_zero: bool = True, activation: Module | None = ReLU(inplace=True))

Overview:

The definition of dynamics network in Stochastic MuZero algorithm, which is used to predict next latent state and reward given current latent state and action.

Parameters:

• num_res_blocks (-) – The number of res blocks in AlphaZero model.

• num_channels (-) – The channels of input, including obs and action encoding.

• reward_head_channels (-) – The channels of reward head.

• fc_reward_layers (-) – The number of hidden layers of the reward head (MLP head).

• output_support_size (-) – The size of categorical reward output.

• flatten_output_size_for_reward_head (-) – The flatten size of output for reward head, i.e., the input size of reward head.

• last_linear_layer_init_zero (-) – Whether to use zero initialization for the last layer of reward mlp, default set it to True.

• activation (-) – Activation function used in network, which often use in-place operation to speedup, e.g. ReLU(inplace=True).

_apply(fn, recurse=True)

_backward_hooks: Dict[int, Callable]

_backward_pre_hooks: Dict[int, Callable]

_buffers: Dict[str, Tensor | None]

_call_impl(*args, **kwargs)

_compiled_call_impl: Callable | None = None

_forward_hooks: Dict[int, Callable]

_forward_hooks_always_called: Dict[int, bool]

_forward_hooks_with_kwargs: Dict[int, bool]

_forward_pre_hooks: Dict[int, Callable]

_forward_pre_hooks_with_kwargs: Dict[int, bool]

_get_backward_hooks()

Return the backward hooks for use in the call function.

It returns two lists, one with the full backward hooks and one with the non-full backward hooks.

_get_backward_pre_hooks()

_get_name()

_is_full_backward_hook: bool | None

_load_from_state_dict(state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs)

Copy parameters and buffers from state_dict into only this module, but not its descendants.

This is called on every submodule in load_state_dict(). Metadata saved for this module in input state_dict is provided as local_metadata. For state dicts without metadata, local_metadata is empty. Subclasses can achieve class-specific backward compatible loading using the version number at local_metadata.get(“version”, None). Additionally, local_metadata can also contain the key assign_to_params_buffers that indicates whether keys should be assigned their corresponding tensor in the state_dict.

Note

state_dict is not the same object as the input state_dict to load_state_dict(). So it can be modified.

Parameters:

• state_dict (dict) – a dict containing parameters and persistent buffers.

• prefix (str) – the prefix for parameters and buffers used in this module

• local_metadata (dict) – a dict containing the metadata for this module. See

• strict (bool) – whether to strictly enforce that the keys in state_dict with prefix match the names of parameters and buffers in this module

• missing_keys (list of str) – if strict=True, add missing keys to this list

• unexpected_keys (list of str) – if strict=True, add unexpected keys to this list

• error_msgs (list of str) – error messages should be added to this list, and will be reported together in load_state_dict()

_load_state_dict_post_hooks: Dict[int, Callable]

_load_state_dict_pre_hooks: Dict[int, Callable]

_maybe_warn_non_full_backward_hook(inputs, result, grad_fn)

_modules: Dict[str, 'Module' | None]

_named_members(get_members_fn, prefix='', recurse=True, remove_duplicate: bool = True)

Help yield various names + members of modules.

_non_persistent_buffers_set: Set[str]

_parameters: Dict[str, Parameter | None]

_register_load_state_dict_pre_hook(hook, with_module=False)

See register_load_state_dict_pre_hook() for details.

A subtle difference is that if with_module is set to False, then the hook will not take the module as the first argument whereas register_load_state_dict_pre_hook() always takes the module as the first argument.

Parameters:

• hook (Callable) – Callable hook that will be invoked before loading the state dict.

• with_module (bool, optional) – Whether or not to pass the module instance to the hook as the first parameter.

_register_state_dict_hook(hook)

Register a post-hook for the state_dict() method.

It should have the following signature::

hook(module, state_dict, prefix, local_metadata) -> None or state_dict

The registered hooks can modify the state_dict inplace or return a new one. If a new state_dict is returned, it will only be respected if it is the root module that state_dict() is called from.

_replicate_for_data_parallel()

_save_to_state_dict(destination, prefix, keep_vars)

Save module state to the destination dictionary.

The destination dictionary will contain the state of the module, but not its descendants. This is called on every submodule in state_dict().

In rare cases, subclasses can achieve class-specific behavior by overriding this method with custom logic.

Parameters:

• destination (dict) – a dict where state will be stored

• prefix (str) – the prefix for parameters and buffers used in this module

_slow_forward(*input, **kwargs)

_state_dict_hooks: Dict[int, Callable]

_state_dict_pre_hooks: Dict[int, Callable]

_version: int = 1

This allows better BC support for load_state_dict(). In state_dict(), the version number will be saved as in the attribute _metadata of the returned state dict, and thus pickled. _metadata is a dictionary 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 shall be bumped, and the module’s _load_from_state_dict method can compare the version number and do appropriate changes if the state dict is from before the change.

_wrapped_call_impl(*args, **kwargs)

add_module(name: str, module: Module | None) → None

Add a child module to the current module.

The module can be accessed as an attribute using the given name.

Parameters:

• name (str) – name of the child module. The child module can be accessed from this module using the given name

• module (Module) – child module to be added to the module.

apply(fn: Callable[[Module], None]) → T

Apply fn recursively to every submodule (as returned by .children()) as well as self.

Typical use includes initializing the parameters of a model (see also nn-init-doc).

Parameters:

fn (Module -> None) – function to be applied to each submodule

Returns:

self

Return type:

Module

Example:

>>> @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_features=2, out_features=2, bias=True)
Parameter containing:
tensor([[1., 1.],
        [1., 1.]], requires_grad=True)
Linear(in_features=2, out_features=2, bias=True)
Parameter containing:
tensor([[1., 1.],
        [1., 1.]], requires_grad=True)
Sequential(
  (0): Linear(in_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
)

bfloat16() → T

Casts all floating point parameters and buffers to bfloat16 datatype.

Note

This method modifies the module in-place.

Returns:

self

Return type:

Module

buffers(recurse: bool = True) → Iterator[Tensor]

Return an iterator over module buffers.

Parameters:

recurse (bool) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module.

Yields:

torch.Tensor – module buffer

Example:

>>> # 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)

call_super_init: bool = False

children() → Iterator[Module]

Return 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

Move all model parameters and buffers to the CPU.

Note

This method modifies the module in-place.

Returns:

self

Return type:

Module

cuda(device: int | device | None = None) → T

Move 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.

Parameters:

device (int, optional) – if specified, all parameters will be copied to that device

Returns:

self

Return type:

Module

double() → T

Casts all floating point parameters and buffers to double datatype.

Note

This method modifies the module in-place.

Returns:

self

Return type:

Module

dump_patches: bool = False

eval() → T

Set 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-disable-grad-doc for a comparison between .eval() and several similar mechanisms that may be confused with it.

Returns:

self

Return type:

Module

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:

Module

forward(state_action_encoding: Tensor) → Tuple[Tensor, Tensor]

Overview:

Forward computation of the dynamics network. Predict next latent state given current latent state and action.

Parameters:

state_action_encoding (-) – The state-action encoding, which is the concatenation of latent state and action encoding, with shape (batch_size, num_channels, height, width).

Returns:

The next latent state, with shape (batch_size, num_channels, height, width).

Return type:

• next_latent_state (torch.Tensor)

• reward (torch.Tensor): The predicted reward, with shape (batch_size, output_support_size).

get_buffer(target: str) → Tensor

Return the buffer given by target if it exists, otherwise throw an error.

See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target.

Parameters:

target – The fully-qualified string name of the buffer to look for. (See get_submodule for how to specify a fully-qualified string.)

Returns:

The buffer referenced by target

Return type:

torch.Tensor

Raises:

AttributeError – If the target string references an invalid path or resolves to something that is not a buffer

get_dynamic_mean() → float

get_extra_state() → Any

Return 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:

object

get_parameter(target: str) → Parameter

Return the parameter given by target if it exists, otherwise throw an error.

See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target.

Parameters:

target – The fully-qualified string name of the Parameter to look for. (See get_submodule for how to specify a fully-qualified string.)

Returns:

The Parameter referenced by target

Return type:

torch.nn.Parameter

Raises:

AttributeError – If the target string references an invalid path or resolves to something that is not an nn.Parameter

get_reward_mean() → float

get_submodule(target: str) → Module

Return the submodule given by target if it exists, otherwise throw 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 submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.)

To check whether or not we have the linear submodule, we would call get_submodule("net_b.linear"). To check whether we have the conv submodule, we would call get_submodule("net_b.net_c.conv").

The runtime of get_submodule is bounded by the degree of module nesting in target. A query against named_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:

torch.nn.Module

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:

Module

ipu(device: int | device | None = None) → T

Move 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.

Parameters:

device (int, optional) – if specified, all parameters will be copied to that device

Returns:

self

Return type:

Module

load_state_dict(state_dict: Mapping[str, Any], strict: bool = True, assign: bool = False)

Copy parameters and buffers from state_dict into this module and its descendants.

If strict is True, then the keys of state_dict must exactly match the keys returned by this module’s state_dict() function.

Warning

If assign is True the optimizer must be created after the call to load_state_dict unless get_swap_module_params_on_conversion() is True.

Parameters:

• state_dict (dict) – a dict containing parameters and persistent buffers.

• strict (bool, optional) – whether to strictly enforce that the keys in state_dict match the keys returned by this module’s state_dict() function. Default: True

• assign (bool, optional) – When False, the properties of the tensors in the current module are preserved while when True, the properties of the Tensors in the state dict are preserved. The only exception is the requires_grad field of Default: ``False`

Returns:

• missing_keys is a list of str containing any keys that are expected

  by this module but missing from the provided state_dict.

• unexpected_keys is a list of str containing the keys that are not

  expected by this module but present in the provided state_dict.

Return type:

NamedTuple with missing_keys and unexpected_keys fields

Note

If a parameter or buffer is registered as None and its corresponding key exists in state_dict, load_state_dict() will raise a RuntimeError.

modules() → Iterator[Module]

Return an iterator over all modules in the network.

Yields:

Module – a module in the network

Note

Duplicate modules are returned only once. In the following example, 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_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
)
1 -> Linear(in_features=2, out_features=2, bias=True)

mtia(device: int | device | None = None) → T

Move all model parameters and buffers to the MTIA.

This also makes associated parameters and buffers different objects. So it should be called before constructing optimizer if the module will live on MTIA while being optimized.

Note

This method modifies the module in-place.

Parameters:

device (int, optional) – if specified, all parameters will be copied to that device

Returns:

self

Return type:

Module

named_buffers(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[Tuple[str, Tensor]]

Return an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.

Parameters:

• prefix (str) – prefix to prepend to all buffer names.

• recurse (bool, optional) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are 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 buffer

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, buf in self.named_buffers():
>>>     if name in ['running_var']:
>>>         print(buf.size())

named_children() → Iterator[Tuple[str, Module]]

Return an iterator over immediate children modules, yielding both the name of the module as well as the module itself.

Yields:

(str, Module) – Tuple containing a name and child module

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, module in model.named_children():
>>>     if name in ['conv4', 'conv5']:
>>>         print(module)

named_modules(memo: Set[Module] | None = None, prefix: str = '', remove_duplicate: bool = True)

Return an iterator over all modules in the network, yielding both the name of the module as well as the module itself.

Parameters:

• memo – a memo to store the set of modules already added to the result

• prefix – a prefix that will be added to the name of the module

• remove_duplicate – whether to remove the duplicated module instances in the result or not

Yields:

(str, Module) – Tuple of name and module

Note

Duplicate modules are returned only once. In the following example, 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_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
))
1 -> ('0', Linear(in_features=2, out_features=2, bias=True))

named_parameters(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[Tuple[str, Parameter]]

Return an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.

Parameters:

• prefix (str) – prefix to prepend to all parameter names.

• recurse (bool) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module.

• remove_duplicate (bool, optional) – whether to remove the duplicated parameters in the result. Defaults to True.

Yields:

(str, Parameter) – Tuple containing the name and parameter

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, param in self.named_parameters():
>>>     if name in ['bias']:
>>>         print(param.size())

parameters(recurse: bool = True) → Iterator[Parameter]

Return an iterator over module parameters.

This is typically passed to an optimizer.

Parameters:

recurse (bool) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module.

Yields:

Parameter – module parameter

Example:

>>> # 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)

register_backward_hook(hook: Callable[[Module, Tuple[Tensor, ...] | Tensor, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor]) → RemovableHandle

Register 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_buffer(name: str, tensor: Tensor | None, persistent: bool = True) → None

Add a buffer to the module.

This is typically used to register a buffer that should not to be considered a model parameter. For example, BatchNorm’s running_mean is not a parameter, but is part of the module’s state. Buffers, by default, are persistent and will be saved alongside parameters. This behavior can be changed by setting persistent to False. The only difference between a persistent buffer and a non-persistent buffer is that the latter will not be a part of this module’s state_dict.

Buffers can be accessed as attributes using given names.

Parameters:

• name (str) – name of the buffer. The buffer can be accessed from this module using the given name

• tensor (Tensor or None) – buffer to be registered. If None, then operations that run on buffers, such as cuda, are ignored. If None, the buffer is not included in the module’s state_dict.

• persistent (bool) – whether the buffer is part of this module’s state_dict.

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> self.register_buffer('running_mean', torch.zeros(num_features))

register_forward_hook(hook: Callable[[T, Tuple[Any, ...], Any], Any | None] | Callable[[T, Tuple[Any, ...], Dict[str, Any], Any], Any | None], *, prepend: bool = False, with_kwargs: bool = False, always_call: bool = False) → RemovableHandle

Register a forward hook on the module.

The hook will be called every time after forward() has computed an output.

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 the forward. The hook can modify the output. It can modify the input inplace but it will not have effect on forward since this is called after forward() is called. The hook should have the following signature:

hook(module, args, output) -> None or modified output

If with_kwargs is True, the forward hook will be passed the kwargs 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 provided hook will be fired before all existing forward hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing forward hooks on this torch.nn.modules.Module. Note that global forward hooks registered with register_module_forward_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

• always_call (bool) – If True the hook 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: Callable[[T, Tuple[Any, ...]], Any | None] | Callable[[T, Tuple[Any, ...], Dict[str, Any]], Tuple[Any, Dict[str, Any]] | None], *, prepend: bool = False, with_kwargs: bool = False) → RemovableHandle

Register 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 the forward. 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 existing forward_pre hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing forward_pre hooks on this torch.nn.modules.Module. Note that global forward_pre hooks registered with register_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, Tuple[Tensor, ...] | Tensor, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle

Register 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 and grad_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 of grad_input in subsequent computations. grad_input will only correspond to the inputs given as positional arguments and all kwarg arguments are ignored. Entries in grad_input and grad_output will be None 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 existing backward hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing backward hooks on this torch.nn.modules.Module. Note that global backward hooks registered with register_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, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle

Register 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 of grad_output in subsequent computations. Entries in grad_output will be None 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 existing backward_pre hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing backward_pre hooks on this torch.nn.modules.Module. Note that global backward_pre hooks registered with register_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)

Register 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 the incompatible_keys argument is a NamedTuple consisting of attributes missing_keys and unexpected_keys. missing_keys is a list of str containing the missing keys and unexpected_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 load_state_dict() with strict=True are affected by modifications the hook makes to missing_keys or unexpected_keys, as expected. Additions to either set of keys will result in an error being thrown when strict=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_load_state_dict_pre_hook(hook)

Register a pre-hook to be run before module’s load_state_dict() is called.

It should have the following signature::

hook(module, state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs) -> None # noqa: B950

Parameters:

hook (Callable) – Callable hook that will be invoked before loading the state dict.

register_module(name: str, module: Module | None) → None

Alias for add_module().

register_parameter(name: str, param: Parameter | None) → None

Add a parameter to the module.

The parameter can be accessed as an attribute using given name.

Parameters:

• name (str) – name of the parameter. The parameter can be accessed from this module using the given name

• param (Parameter or None) – parameter to be added to the module. If None, then operations that run on parameters, such as cuda, are ignored. If None, the parameter is not included in the module’s state_dict.

register_state_dict_post_hook(hook)

Register a post-hook for the state_dict() method.

It should have the following signature::

hook(module, state_dict, prefix, local_metadata) -> None

The registered hooks can modify the state_dict inplace.

register_state_dict_pre_hook(hook)

Register a pre-hook for the state_dict() method.

It should have the following signature::

hook(module, prefix, keep_vars) -> None

The registered hooks can be used to perform pre-processing before the state_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-disable-grad-doc for a comparison between .requires_grad_() and several similar mechanisms that may be confused with it.

Parameters:

requires_grad (bool) – whether autograd should record operations on parameters in this module. Default: True.

Returns:

self

Return type:

Module

set_extra_state(state: Any) → None

Set extra state contained in the loaded state_dict.

This function is called from load_state_dict() to handle any extra state found within the state_dict. Implement this function and a corresponding get_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

set_submodule(target: str, module: Module) → None

Set the submodule given by target if it exists, otherwise throw 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 submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.)

To overide the Conv2d with a new submodule Linear, you would call set_submodule("net_b.net_c.conv", nn.Linear(33, 16)).

Parameters:

• target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.)

• module – The module to set the submodule to.

Raises:

• ValueError – If the target string is empty

• AttributeError – If the target string references an invalid path or resolves to something that is not an nn.Module

share_memory() → T

See torch.Tensor.share_memory_().

state_dict(*args, destination=None, prefix='', keep_vars=False)

Return a dictionary containing references to the whole state of the module.

Both parameters and persistent buffers (e.g. running averages) are included. 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 references to the module’s parameters and buffers.

Warning

Currently state_dict() also accepts positional arguments for destination, prefix and keep_vars in order. However, this is being deprecated and keyword arguments will be enforced in future releases.

Warning

Please avoid the use of argument destination as it is not designed for end-users.

Parameters:

• destination (dict, optional) – If provided, the state of module will be 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 buffer names to compose the keys in state_dict. Default: ''.

• keep_vars (bool, optional) – by default the Tensor s returned in the state dict are detached from autograd. If it’s set to True, detaching will not be performed. Default: False.

Returns:

a dictionary containing a whole state of the module

Return type:

dict

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> module.state_dict().keys()
['bias', 'weight']

to(*args, **kwargs)

Move and/or cast 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 complex dtypes. In addition, this method will only cast the floating point or complex parameters and buffers to dtype (if given). The integral parameters and buffers will be moved device, if that is given, but with dtypes unchanged. When non_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 module

• dtype (torch.dtype) – the desired floating point or complex dtype of the parameters and buffers in this module

• tensor (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:

Module

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: int | str | device | None, recurse: bool = True) → T

Move 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:

Module

train(mode: bool = True) → T

Set the module in training 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.

Parameters:

mode (bool) – whether to set training mode (True) or evaluation mode (False). Default: True.

Returns:

self

Return type:

Module

training: bool

type(dst_type: dtype | str) → T

Casts all parameters and buffers to dst_type.

Note

This method modifies the module in-place.

Parameters:

dst_type (type or string) – the desired type

Returns:

self

Return type:

Module

xpu(device: int | device | None = None) → T

Move 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.

Parameters:

device (int, optional) – if specified, all parameters will be copied to that device

Returns:

self

Return type:

Module

zero_grad(set_to_none: bool = True) → None

Reset 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.

class lzero.model.stochastic_muzero_model.AfterstatePredictionNetwork(action_space_size: int, num_res_blocks: int, num_channels: int, value_head_channels: int, policy_head_channels: int, fc_value_layers: int, fc_policy_layers: int, output_support_size: int, flatten_output_size_for_value_head: int, flatten_output_size_for_policy_head: int, last_linear_layer_init_zero: bool = True, activation: Module = ReLU(inplace=True))

Bases: Module

T_destination = ~T_destination

__init__(action_space_size: int, num_res_blocks: int, num_channels: int, value_head_channels: int, policy_head_channels: int, fc_value_layers: int, fc_policy_layers: int, output_support_size: int, flatten_output_size_for_value_head: int, flatten_output_size_for_policy_head: int, last_linear_layer_init_zero: bool = True, activation: Module = ReLU(inplace=True)) → None

Overview:

The definition of afterstate policy and value prediction network, which is used to predict value and policy by the given afterstate.

Parameters:

• action_space_size (-) – (int): Action space size, usually an integer number for discrete action space.

• num_res_blocks (-) – The number of res blocks in AlphaZero model.

• num_channels (-) – The channels of hidden states.

• value_head_channels (-) – The channels of value head.

• policy_head_channels (-) – The channels of policy head.

• fc_value_layers (-) – The number of hidden layers used in value head (MLP head).

• fc_policy_layers (-) – The number of hidden layers used in policy head (MLP head).

• output_support_size (-) – The size of categorical value output.

• self_supervised_learning_loss (-) – Whether to use self_supervised_learning related networks - flatten_output_size_for_value_head (int): The size of flatten hidden states, i.e. the input size of the value head.

• flatten_output_size_for_policy_head (-) – The size of flatten hidden states, i.e. the input size of the policy head.

• last_linear_layer_init_zero (-) – Whether to use zero initialization for the last layer of dynamics/prediction mlp, default set it to True.

• activation (-) – Activation function used in network, which often use in-place operation to speedup, e.g. ReLU(inplace=True).

_apply(fn, recurse=True)

_backward_hooks: Dict[int, Callable]

_backward_pre_hooks: Dict[int, Callable]

_buffers: Dict[str, Tensor | None]

_call_impl(*args, **kwargs)

_compiled_call_impl: Callable | None = None

_forward_hooks: Dict[int, Callable]

_forward_hooks_always_called: Dict[int, bool]

_forward_hooks_with_kwargs: Dict[int, bool]

_forward_pre_hooks: Dict[int, Callable]

_forward_pre_hooks_with_kwargs: Dict[int, bool]

_get_backward_hooks()

Return the backward hooks for use in the call function.

It returns two lists, one with the full backward hooks and one with the non-full backward hooks.

_get_backward_pre_hooks()

_get_name()

_is_full_backward_hook: bool | None

_load_from_state_dict(state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs)

Copy parameters and buffers from state_dict into only this module, but not its descendants.

This is called on every submodule in load_state_dict(). Metadata saved for this module in input state_dict is provided as local_metadata. For state dicts without metadata, local_metadata is empty. Subclasses can achieve class-specific backward compatible loading using the version number at local_metadata.get(“version”, None). Additionally, local_metadata can also contain the key assign_to_params_buffers that indicates whether keys should be assigned their corresponding tensor in the state_dict.

Note

state_dict is not the same object as the input state_dict to load_state_dict(). So it can be modified.

Parameters:

• state_dict (dict) – a dict containing parameters and persistent buffers.

• prefix (str) – the prefix for parameters and buffers used in this module

• local_metadata (dict) – a dict containing the metadata for this module. See

• strict (bool) – whether to strictly enforce that the keys in state_dict with prefix match the names of parameters and buffers in this module

• missing_keys (list of str) – if strict=True, add missing keys to this list

• unexpected_keys (list of str) – if strict=True, add unexpected keys to this list

• error_msgs (list of str) – error messages should be added to this list, and will be reported together in load_state_dict()

_load_state_dict_post_hooks: Dict[int, Callable]

_load_state_dict_pre_hooks: Dict[int, Callable]

_maybe_warn_non_full_backward_hook(inputs, result, grad_fn)

_modules: Dict[str, 'Module' | None]

_named_members(get_members_fn, prefix='', recurse=True, remove_duplicate: bool = True)

Help yield various names + members of modules.

_non_persistent_buffers_set: Set[str]

_parameters: Dict[str, Parameter | None]

_register_load_state_dict_pre_hook(hook, with_module=False)

See register_load_state_dict_pre_hook() for details.

A subtle difference is that if with_module is set to False, then the hook will not take the module as the first argument whereas register_load_state_dict_pre_hook() always takes the module as the first argument.

Parameters:

• hook (Callable) – Callable hook that will be invoked before loading the state dict.

• with_module (bool, optional) – Whether or not to pass the module instance to the hook as the first parameter.

_register_state_dict_hook(hook)

Register a post-hook for the state_dict() method.

It should have the following signature::

hook(module, state_dict, prefix, local_metadata) -> None or state_dict

The registered hooks can modify the state_dict inplace or return a new one. If a new state_dict is returned, it will only be respected if it is the root module that state_dict() is called from.

_replicate_for_data_parallel()

_save_to_state_dict(destination, prefix, keep_vars)

Save module state to the destination dictionary.

The destination dictionary will contain the state of the module, but not its descendants. This is called on every submodule in state_dict().

In rare cases, subclasses can achieve class-specific behavior by overriding this method with custom logic.

Parameters:

• destination (dict) – a dict where state will be stored

• prefix (str) – the prefix for parameters and buffers used in this module

_slow_forward(*input, **kwargs)

_state_dict_hooks: Dict[int, Callable]

_state_dict_pre_hooks: Dict[int, Callable]

_version: int = 1

This allows better BC support for load_state_dict(). In state_dict(), the version number will be saved as in the attribute _metadata of the returned state dict, and thus pickled. _metadata is a dictionary 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 shall be bumped, and the module’s _load_from_state_dict method can compare the version number and do appropriate changes if the state dict is from before the change.

_wrapped_call_impl(*args, **kwargs)

add_module(name: str, module: Module | None) → None

Add a child module to the current module.

The module can be accessed as an attribute using the given name.

Parameters:

• name (str) – name of the child module. The child module can be accessed from this module using the given name

• module (Module) – child module to be added to the module.

apply(fn: Callable[[Module], None]) → T

Apply fn recursively to every submodule (as returned by .children()) as well as self.

Typical use includes initializing the parameters of a model (see also nn-init-doc).

Parameters:

fn (Module -> None) – function to be applied to each submodule

Returns:

self

Return type:

Module

Example:

>>> @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_features=2, out_features=2, bias=True)
Parameter containing:
tensor([[1., 1.],
        [1., 1.]], requires_grad=True)
Linear(in_features=2, out_features=2, bias=True)
Parameter containing:
tensor([[1., 1.],
        [1., 1.]], requires_grad=True)
Sequential(
  (0): Linear(in_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
)

bfloat16() → T

Casts all floating point parameters and buffers to bfloat16 datatype.

Note

This method modifies the module in-place.

Returns:

self

Return type:

Module

buffers(recurse: bool = True) → Iterator[Tensor]

Return an iterator over module buffers.

Parameters:

recurse (bool) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module.

Yields:

torch.Tensor – module buffer

Example:

>>> # 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)

call_super_init: bool = False

children() → Iterator[Module]

Return 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

Move all model parameters and buffers to the CPU.

Note

This method modifies the module in-place.

Returns:

self

Return type:

Module

cuda(device: int | device | None = None) → T

Move 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.

Parameters:

device (int, optional) – if specified, all parameters will be copied to that device

Returns:

self

Return type:

Module

double() → T

Casts all floating point parameters and buffers to double datatype.

Note

This method modifies the module in-place.

Returns:

self

Return type:

Module

dump_patches: bool = False

eval() → T

Set 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-disable-grad-doc for a comparison between .eval() and several similar mechanisms that may be confused with it.

Returns:

self

Return type:

Module

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:

Module

forward(afterstate: Tensor) → Tuple[Tensor, Tensor]

Overview:

Forward computation of the afterstate prediction network.

Parameters:

afterstate (-) – input tensor with shape (B, afterstate_dim).

Returns:

policy tensor with shape (B, action_space_size). - afterstate_value (torch.Tensor): value tensor with shape (B, output_support_size).

Return type:

• afterstate_policy_logits (torch.Tensor)

get_buffer(target: str) → Tensor

Return the buffer given by target if it exists, otherwise throw an error.

See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target.

Parameters:

target – The fully-qualified string name of the buffer to look for. (See get_submodule for how to specify a fully-qualified string.)

Returns:

The buffer referenced by target

Return type:

torch.Tensor

Raises:

AttributeError – If the target string references an invalid path or resolves to something that is not a buffer

get_extra_state() → Any

Return 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:

object

get_parameter(target: str) → Parameter

Return the parameter given by target if it exists, otherwise throw an error.

See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target.

Parameters:

target – The fully-qualified string name of the Parameter to look for. (See get_submodule for how to specify a fully-qualified string.)

Returns:

The Parameter referenced by target

Return type:

torch.nn.Parameter

Raises:

AttributeError – If the target string references an invalid path or resolves to something that is not an nn.Parameter

get_submodule(target: str) → Module

Return the submodule given by target if it exists, otherwise throw 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 submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.)

To check whether or not we have the linear submodule, we would call get_submodule("net_b.linear"). To check whether we have the conv submodule, we would call get_submodule("net_b.net_c.conv").

The runtime of get_submodule is bounded by the degree of module nesting in target. A query against named_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:

torch.nn.Module

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:

Module

ipu(device: int | device | None = None) → T

Move 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.

Parameters:

device (int, optional) – if specified, all parameters will be copied to that device

Returns:

self

Return type:

Module

load_state_dict(state_dict: Mapping[str, Any], strict: bool = True, assign: bool = False)

Copy parameters and buffers from state_dict into this module and its descendants.

If strict is True, then the keys of state_dict must exactly match the keys returned by this module’s state_dict() function.

Warning

If assign is True the optimizer must be created after the call to load_state_dict unless get_swap_module_params_on_conversion() is True.

Parameters:

• state_dict (dict) – a dict containing parameters and persistent buffers.

• strict (bool, optional) – whether to strictly enforce that the keys in state_dict match the keys returned by this module’s state_dict() function. Default: True

• assign (bool, optional) – When False, the properties of the tensors in the current module are preserved while when True, the properties of the Tensors in the state dict are preserved. The only exception is the requires_grad field of Default: ``False`

Returns:

• missing_keys is a list of str containing any keys that are expected

  by this module but missing from the provided state_dict.

• unexpected_keys is a list of str containing the keys that are not

  expected by this module but present in the provided state_dict.

Return type:

NamedTuple with missing_keys and unexpected_keys fields

Note

If a parameter or buffer is registered as None and its corresponding key exists in state_dict, load_state_dict() will raise a RuntimeError.

modules() → Iterator[Module]

Return an iterator over all modules in the network.

Yields:

Module – a module in the network

Note

Duplicate modules are returned only once. In the following example, 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_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
)
1 -> Linear(in_features=2, out_features=2, bias=True)

mtia(device: int | device | None = None) → T

Move all model parameters and buffers to the MTIA.

This also makes associated parameters and buffers different objects. So it should be called before constructing optimizer if the module will live on MTIA while being optimized.

Note

This method modifies the module in-place.

Parameters:

device (int, optional) – if specified, all parameters will be copied to that device

Returns:

self

Return type:

Module

named_buffers(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[Tuple[str, Tensor]]

Return an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.

Parameters:

• prefix (str) – prefix to prepend to all buffer names.

• recurse (bool, optional) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are 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 buffer

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, buf in self.named_buffers():
>>>     if name in ['running_var']:
>>>         print(buf.size())

named_children() → Iterator[Tuple[str, Module]]

Return an iterator over immediate children modules, yielding both the name of the module as well as the module itself.

Yields:

(str, Module) – Tuple containing a name and child module

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, module in model.named_children():
>>>     if name in ['conv4', 'conv5']:
>>>         print(module)

named_modules(memo: Set[Module] | None = None, prefix: str = '', remove_duplicate: bool = True)

Return an iterator over all modules in the network, yielding both the name of the module as well as the module itself.

Parameters:

• memo – a memo to store the set of modules already added to the result

• prefix – a prefix that will be added to the name of the module

• remove_duplicate – whether to remove the duplicated module instances in the result or not

Yields:

(str, Module) – Tuple of name and module

Note

Duplicate modules are returned only once. In the following example, 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_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
))
1 -> ('0', Linear(in_features=2, out_features=2, bias=True))

named_parameters(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[Tuple[str, Parameter]]

Return an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.

Parameters:

• prefix (str) – prefix to prepend to all parameter names.

• recurse (bool) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module.

• remove_duplicate (bool, optional) – whether to remove the duplicated parameters in the result. Defaults to True.

Yields:

(str, Parameter) – Tuple containing the name and parameter

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, param in self.named_parameters():
>>>     if name in ['bias']:
>>>         print(param.size())

parameters(recurse: bool = True) → Iterator[Parameter]

Return an iterator over module parameters.

This is typically passed to an optimizer.

Parameters:

recurse (bool) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module.

Yields:

Parameter – module parameter

Example:

>>> # 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)

register_backward_hook(hook: Callable[[Module, Tuple[Tensor, ...] | Tensor, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor]) → RemovableHandle

Register 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_buffer(name: str, tensor: Tensor | None, persistent: bool = True) → None

Add a buffer to the module.

This is typically used to register a buffer that should not to be considered a model parameter. For example, BatchNorm’s running_mean is not a parameter, but is part of the module’s state. Buffers, by default, are persistent and will be saved alongside parameters. This behavior can be changed by setting persistent to False. The only difference between a persistent buffer and a non-persistent buffer is that the latter will not be a part of this module’s state_dict.

Buffers can be accessed as attributes using given names.

Parameters:

• name (str) – name of the buffer. The buffer can be accessed from this module using the given name

• tensor (Tensor or None) – buffer to be registered. If None, then operations that run on buffers, such as cuda, are ignored. If None, the buffer is not included in the module’s state_dict.

• persistent (bool) – whether the buffer is part of this module’s state_dict.

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> self.register_buffer('running_mean', torch.zeros(num_features))

register_forward_hook(hook: Callable[[T, Tuple[Any, ...], Any], Any | None] | Callable[[T, Tuple[Any, ...], Dict[str, Any], Any], Any | None], *, prepend: bool = False, with_kwargs: bool = False, always_call: bool = False) → RemovableHandle

Register a forward hook on the module.

The hook will be called every time after forward() has computed an output.

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 the forward. The hook can modify the output. It can modify the input inplace but it will not have effect on forward since this is called after forward() is called. The hook should have the following signature:

hook(module, args, output) -> None or modified output

If with_kwargs is True, the forward hook will be passed the kwargs 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 provided hook will be fired before all existing forward hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing forward hooks on this torch.nn.modules.Module. Note that global forward hooks registered with register_module_forward_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

• always_call (bool) – If True the hook 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: Callable[[T, Tuple[Any, ...]], Any | None] | Callable[[T, Tuple[Any, ...], Dict[str, Any]], Tuple[Any, Dict[str, Any]] | None], *, prepend: bool = False, with_kwargs: bool = False) → RemovableHandle

Register 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 the forward. 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 existing forward_pre hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing forward_pre hooks on this torch.nn.modules.Module. Note that global forward_pre hooks registered with register_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, Tuple[Tensor, ...] | Tensor, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle

Register 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 and grad_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 of grad_input in subsequent computations. grad_input will only correspond to the inputs given as positional arguments and all kwarg arguments are ignored. Entries in grad_input and grad_output will be None 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 existing backward hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing backward hooks on this torch.nn.modules.Module. Note that global backward hooks registered with register_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, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle

Register 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 of grad_output in subsequent computations. Entries in grad_output will be None 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 existing backward_pre hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing backward_pre hooks on this torch.nn.modules.Module. Note that global backward_pre hooks registered with register_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)

Register 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 the incompatible_keys argument is a NamedTuple consisting of attributes missing_keys and unexpected_keys. missing_keys is a list of str containing the missing keys and unexpected_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 load_state_dict() with strict=True are affected by modifications the hook makes to missing_keys or unexpected_keys, as expected. Additions to either set of keys will result in an error being thrown when strict=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_load_state_dict_pre_hook(hook)

Register a pre-hook to be run before module’s load_state_dict() is called.

It should have the following signature::

hook(module, state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs) -> None # noqa: B950

Parameters:

hook (Callable) – Callable hook that will be invoked before loading the state dict.

register_module(name: str, module: Module | None) → None

Alias for add_module().

register_parameter(name: str, param: Parameter | None) → None

Add a parameter to the module.

The parameter can be accessed as an attribute using given name.

Parameters:

• name (str) – name of the parameter. The parameter can be accessed from this module using the given name

• param (Parameter or None) – parameter to be added to the module. If None, then operations that run on parameters, such as cuda, are ignored. If None, the parameter is not included in the module’s state_dict.

register_state_dict_post_hook(hook)

Register a post-hook for the state_dict() method.

It should have the following signature::

hook(module, state_dict, prefix, local_metadata) -> None

The registered hooks can modify the state_dict inplace.

register_state_dict_pre_hook(hook)

Register a pre-hook for the state_dict() method.

It should have the following signature::

hook(module, prefix, keep_vars) -> None

The registered hooks can be used to perform pre-processing before the state_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-disable-grad-doc for a comparison between .requires_grad_() and several similar mechanisms that may be confused with it.

Parameters:

requires_grad (bool) – whether autograd should record operations on parameters in this module. Default: True.

Returns:

self

Return type:

Module

set_extra_state(state: Any) → None

Set extra state contained in the loaded state_dict.

This function is called from load_state_dict() to handle any extra state found within the state_dict. Implement this function and a corresponding get_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

set_submodule(target: str, module: Module) → None

Set the submodule given by target if it exists, otherwise throw 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 submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.)

To overide the Conv2d with a new submodule Linear, you would call set_submodule("net_b.net_c.conv", nn.Linear(33, 16)).

Parameters:

• target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.)

• module – The module to set the submodule to.

Raises:

• ValueError – If the target string is empty

• AttributeError – If the target string references an invalid path or resolves to something that is not an nn.Module

share_memory() → T

See torch.Tensor.share_memory_().

state_dict(*args, destination=None, prefix='', keep_vars=False)

Return a dictionary containing references to the whole state of the module.

Both parameters and persistent buffers (e.g. running averages) are included. 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 references to the module’s parameters and buffers.

Warning

Currently state_dict() also accepts positional arguments for destination, prefix and keep_vars in order. However, this is being deprecated and keyword arguments will be enforced in future releases.

Warning

Please avoid the use of argument destination as it is not designed for end-users.

Parameters:

• destination (dict, optional) – If provided, the state of module will be 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 buffer names to compose the keys in state_dict. Default: ''.

• keep_vars (bool, optional) – by default the Tensor s returned in the state dict are detached from autograd. If it’s set to True, detaching will not be performed. Default: False.

Returns:

a dictionary containing a whole state of the module

Return type:

dict

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> module.state_dict().keys()
['bias', 'weight']

to(*args, **kwargs)

Move and/or cast 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 complex dtypes. In addition, this method will only cast the floating point or complex parameters and buffers to dtype (if given). The integral parameters and buffers will be moved device, if that is given, but with dtypes unchanged. When non_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 module

• dtype (torch.dtype) – the desired floating point or complex dtype of the parameters and buffers in this module

• tensor (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:

Module

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: int | str | device | None, recurse: bool = True) → T

Move 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:

Module

train(mode: bool = True) → T

Set the module in training 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.

Parameters:

mode (bool) – whether to set training mode (True) or evaluation mode (False). Default: True.

Returns:

self

Return type:

Module

training: bool

type(dst_type: dtype | str) → T

Casts all parameters and buffers to dst_type.

Note

This method modifies the module in-place.

Parameters:

dst_type (type or string) – the desired type

Returns:

self

Return type:

Module

xpu(device: int | device | None = None) → T

Move 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.

Parameters:

device (int, optional) – if specified, all parameters will be copied to that device

Returns:

self

Return type:

Module

zero_grad(set_to_none: bool = True) → None

Reset 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.

class lzero.model.stochastic_muzero_model.ChanceEncoderBackbone(input_dimensions, chance_encoding_dim=4)

Bases: Module

Overview:

The definition of chance encoder backbone network, which is used to encode the (image) observation into a latent space.

Parameters:

• input_dimensions (-) – The dimension of observation space.

• chance_encoding_dim (-) – The dimension of chance encoding.

T_destination = ~T_destination

__init__(input_dimensions, chance_encoding_dim=4)

Initialize internal Module state, shared by both nn.Module and ScriptModule.

_apply(fn, recurse=True)

_backward_hooks: Dict[int, Callable]

_backward_pre_hooks: Dict[int, Callable]

_buffers: Dict[str, Tensor | None]

_call_impl(*args, **kwargs)

_compiled_call_impl: Callable | None = None

_forward_hooks: Dict[int, Callable]

_forward_hooks_always_called: Dict[int, bool]

_forward_hooks_with_kwargs: Dict[int, bool]

_forward_pre_hooks: Dict[int, Callable]

_forward_pre_hooks_with_kwargs: Dict[int, bool]

_get_backward_hooks()

Return the backward hooks for use in the call function.

It returns two lists, one with the full backward hooks and one with the non-full backward hooks.

_get_backward_pre_hooks()

_get_name()

_is_full_backward_hook: bool | None

_load_from_state_dict(state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs)

Copy parameters and buffers from state_dict into only this module, but not its descendants.

This is called on every submodule in load_state_dict(). Metadata saved for this module in input state_dict is provided as local_metadata. For state dicts without metadata, local_metadata is empty. Subclasses can achieve class-specific backward compatible loading using the version number at local_metadata.get(“version”, None). Additionally, local_metadata can also contain the key assign_to_params_buffers that indicates whether keys should be assigned their corresponding tensor in the state_dict.

Note

state_dict is not the same object as the input state_dict to load_state_dict(). So it can be modified.

Parameters:

• state_dict (dict) – a dict containing parameters and persistent buffers.

• prefix (str) – the prefix for parameters and buffers used in this module

• local_metadata (dict) – a dict containing the metadata for this module. See

• strict (bool) – whether to strictly enforce that the keys in state_dict with prefix match the names of parameters and buffers in this module

• missing_keys (list of str) – if strict=True, add missing keys to this list

• unexpected_keys (list of str) – if strict=True, add unexpected keys to this list

• error_msgs (list of str) – error messages should be added to this list, and will be reported together in load_state_dict()

_load_state_dict_post_hooks: Dict[int, Callable]

_load_state_dict_pre_hooks: Dict[int, Callable]

_maybe_warn_non_full_backward_hook(inputs, result, grad_fn)

_modules: Dict[str, 'Module' | None]

_named_members(get_members_fn, prefix='', recurse=True, remove_duplicate: bool = True)

Help yield various names + members of modules.

_non_persistent_buffers_set: Set[str]

_parameters: Dict[str, Parameter | None]

_register_load_state_dict_pre_hook(hook, with_module=False)

See register_load_state_dict_pre_hook() for details.

A subtle difference is that if with_module is set to False, then the hook will not take the module as the first argument whereas register_load_state_dict_pre_hook() always takes the module as the first argument.

Parameters:

• hook (Callable) – Callable hook that will be invoked before loading the state dict.

• with_module (bool, optional) – Whether or not to pass the module instance to the hook as the first parameter.

_register_state_dict_hook(hook)

Register a post-hook for the state_dict() method.

It should have the following signature::

hook(module, state_dict, prefix, local_metadata) -> None or state_dict

The registered hooks can modify the state_dict inplace or return a new one. If a new state_dict is returned, it will only be respected if it is the root module that state_dict() is called from.

_replicate_for_data_parallel()

_save_to_state_dict(destination, prefix, keep_vars)

Save module state to the destination dictionary.

The destination dictionary will contain the state of the module, but not its descendants. This is called on every submodule in state_dict().

In rare cases, subclasses can achieve class-specific behavior by overriding this method with custom logic.

Parameters:

• destination (dict) – a dict where state will be stored

• prefix (str) – the prefix for parameters and buffers used in this module

_slow_forward(*input, **kwargs)

_state_dict_hooks: Dict[int, Callable]

_state_dict_pre_hooks: Dict[int, Callable]

_version: int = 1

This allows better BC support for load_state_dict(). In state_dict(), the version number will be saved as in the attribute _metadata of the returned state dict, and thus pickled. _metadata is a dictionary 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 shall be bumped, and the module’s _load_from_state_dict method can compare the version number and do appropriate changes if the state dict is from before the change.

_wrapped_call_impl(*args, **kwargs)

add_module(name: str, module: Module | None) → None

Add a child module to the current module.

The module can be accessed as an attribute using the given name.

Parameters:

• name (str) – name of the child module. The child module can be accessed from this module using the given name

• module (Module) – child module to be added to the module.

apply(fn: Callable[[Module], None]) → T

Apply fn recursively to every submodule (as returned by .children()) as well as self.

Typical use includes initializing the parameters of a model (see also nn-init-doc).

Parameters:

fn (Module -> None) – function to be applied to each submodule

Returns:

self

Return type:

Module

Example:

>>> @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_features=2, out_features=2, bias=True)
Parameter containing:
tensor([[1., 1.],
        [1., 1.]], requires_grad=True)
Linear(in_features=2, out_features=2, bias=True)
Parameter containing:
tensor([[1., 1.],
        [1., 1.]], requires_grad=True)
Sequential(
  (0): Linear(in_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
)

bfloat16() → T

Casts all floating point parameters and buffers to bfloat16 datatype.

Note

This method modifies the module in-place.

Returns:

self

Return type:

Module

buffers(recurse: bool = True) → Iterator[Tensor]

Return an iterator over module buffers.

Parameters:

recurse (bool) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module.

Yields:

torch.Tensor – module buffer

Example:

>>> # 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)

call_super_init: bool = False

children() → Iterator[Module]

Return 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

Move all model parameters and buffers to the CPU.

Note

This method modifies the module in-place.

Returns:

self

Return type:

Module

cuda(device: int | device | None = None) → T

Move 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.

Parameters:

device (int, optional) – if specified, all parameters will be copied to that device

Returns:

self

Return type:

Module

double() → T

Casts all floating point parameters and buffers to double datatype.

Note

This method modifies the module in-place.

Returns:

self

Return type:

Module

dump_patches: bool = False

eval() → T

Set 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-disable-grad-doc for a comparison between .eval() and several similar mechanisms that may be confused with it.

Returns:

self

Return type:

Module

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:

Module

forward(x)

Define the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

get_buffer(target: str) → Tensor

Return the buffer given by target if it exists, otherwise throw an error.

See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target.

Parameters:

target – The fully-qualified string name of the buffer to look for. (See get_submodule for how to specify a fully-qualified string.)

Returns:

The buffer referenced by target

Return type:

torch.Tensor

Raises:

AttributeError – If the target string references an invalid path or resolves to something that is not a buffer

get_extra_state() → Any

Return 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:

object

get_parameter(target: str) → Parameter

Return the parameter given by target if it exists, otherwise throw an error.

See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target.

Parameters:

target – The fully-qualified string name of the Parameter to look for. (See get_submodule for how to specify a fully-qualified string.)

Returns:

The Parameter referenced by target

Return type:

torch.nn.Parameter

Raises:

AttributeError – If the target string references an invalid path or resolves to something that is not an nn.Parameter

get_submodule(target: str) → Module

Return the submodule given by target if it exists, otherwise throw 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 submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.)

To check whether or not we have the linear submodule, we would call get_submodule("net_b.linear"). To check whether we have the conv submodule, we would call get_submodule("net_b.net_c.conv").

The runtime of get_submodule is bounded by the degree of module nesting in target. A query against named_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:

torch.nn.Module

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:

Module

ipu(device: int | device | None = None) → T

Move 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.

Parameters:

device (int, optional) – if specified, all parameters will be copied to that device

Returns:

self

Return type:

Module

load_state_dict(state_dict: Mapping[str, Any], strict: bool = True, assign: bool = False)

Copy parameters and buffers from state_dict into this module and its descendants.

If strict is True, then the keys of state_dict must exactly match the keys returned by this module’s state_dict() function.

Warning

If assign is True the optimizer must be created after the call to load_state_dict unless get_swap_module_params_on_conversion() is True.

Parameters:

• state_dict (dict) – a dict containing parameters and persistent buffers.

• strict (bool, optional) – whether to strictly enforce that the keys in state_dict match the keys returned by this module’s state_dict() function. Default: True

• assign (bool, optional) – When False, the properties of the tensors in the current module are preserved while when True, the properties of the Tensors in the state dict are preserved. The only exception is the requires_grad field of Default: ``False`

Returns:

• missing_keys is a list of str containing any keys that are expected

  by this module but missing from the provided state_dict.

• unexpected_keys is a list of str containing the keys that are not

  expected by this module but present in the provided state_dict.

Return type:

NamedTuple with missing_keys and unexpected_keys fields

Note

If a parameter or buffer is registered as None and its corresponding key exists in state_dict, load_state_dict() will raise a RuntimeError.

modules() → Iterator[Module]

Return an iterator over all modules in the network.

Yields:

Module – a module in the network

Note

Duplicate modules are returned only once. In the following example, 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_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
)
1 -> Linear(in_features=2, out_features=2, bias=True)

mtia(device: int | device | None = None) → T

Move all model parameters and buffers to the MTIA.

This also makes associated parameters and buffers different objects. So it should be called before constructing optimizer if the module will live on MTIA while being optimized.

Note

This method modifies the module in-place.

Parameters:

device (int, optional) – if specified, all parameters will be copied to that device

Returns:

self

Return type:

Module

named_buffers(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[Tuple[str, Tensor]]

Return an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.

Parameters:

• prefix (str) – prefix to prepend to all buffer names.

• recurse (bool, optional) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are 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 buffer

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, buf in self.named_buffers():
>>>     if name in ['running_var']:
>>>         print(buf.size())

named_children() → Iterator[Tuple[str, Module]]

Return an iterator over immediate children modules, yielding both the name of the module as well as the module itself.

Yields:

(str, Module) – Tuple containing a name and child module

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, module in model.named_children():
>>>     if name in ['conv4', 'conv5']:
>>>         print(module)

named_modules(memo: Set[Module] | None = None, prefix: str = '', remove_duplicate: bool = True)

Return an iterator over all modules in the network, yielding both the name of the module as well as the module itself.

Parameters:

• memo – a memo to store the set of modules already added to the result

• prefix – a prefix that will be added to the name of the module

• remove_duplicate – whether to remove the duplicated module instances in the result or not

Yields:

(str, Module) – Tuple of name and module

Note

Duplicate modules are returned only once. In the following example, 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_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
))
1 -> ('0', Linear(in_features=2, out_features=2, bias=True))

named_parameters(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[Tuple[str, Parameter]]

Return an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.

Parameters:

• prefix (str) – prefix to prepend to all parameter names.

• recurse (bool) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module.

• remove_duplicate (bool, optional) – whether to remove the duplicated parameters in the result. Defaults to True.

Yields:

(str, Parameter) – Tuple containing the name and parameter

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, param in self.named_parameters():
>>>     if name in ['bias']:
>>>         print(param.size())

parameters(recurse: bool = True) → Iterator[Parameter]

Return an iterator over module parameters.

This is typically passed to an optimizer.

Parameters:

recurse (bool) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module.

Yields:

Parameter – module parameter

Example:

>>> # 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)

register_backward_hook(hook: Callable[[Module, Tuple[Tensor, ...] | Tensor, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor]) → RemovableHandle

Register 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_buffer(name: str, tensor: Tensor | None, persistent: bool = True) → None

Add a buffer to the module.

This is typically used to register a buffer that should not to be considered a model parameter. For example, BatchNorm’s running_mean is not a parameter, but is part of the module’s state. Buffers, by default, are persistent and will be saved alongside parameters. This behavior can be changed by setting persistent to False. The only difference between a persistent buffer and a non-persistent buffer is that the latter will not be a part of this module’s state_dict.

Buffers can be accessed as attributes using given names.

Parameters:

• name (str) – name of the buffer. The buffer can be accessed from this module using the given name

• tensor (Tensor or None) – buffer to be registered. If None, then operations that run on buffers, such as cuda, are ignored. If None, the buffer is not included in the module’s state_dict.

• persistent (bool) – whether the buffer is part of this module’s state_dict.

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> self.register_buffer('running_mean', torch.zeros(num_features))

register_forward_hook(hook: Callable[[T, Tuple[Any, ...], Any], Any | None] | Callable[[T, Tuple[Any, ...], Dict[str, Any], Any], Any | None], *, prepend: bool = False, with_kwargs: bool = False, always_call: bool = False) → RemovableHandle

Register a forward hook on the module.

The hook will be called every time after forward() has computed an output.

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 the forward. The hook can modify the output. It can modify the input inplace but it will not have effect on forward since this is called after forward() is called. The hook should have the following signature:

hook(module, args, output) -> None or modified output

If with_kwargs is True, the forward hook will be passed the kwargs 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 provided hook will be fired before all existing forward hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing forward hooks on this torch.nn.modules.Module. Note that global forward hooks registered with register_module_forward_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

• always_call (bool) – If True the hook 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: Callable[[T, Tuple[Any, ...]], Any | None] | Callable[[T, Tuple[Any, ...], Dict[str, Any]], Tuple[Any, Dict[str, Any]] | None], *, prepend: bool = False, with_kwargs: bool = False) → RemovableHandle

Register 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 the forward. 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 existing forward_pre hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing forward_pre hooks on this torch.nn.modules.Module. Note that global forward_pre hooks registered with register_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, Tuple[Tensor, ...] | Tensor, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle

Register 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 and grad_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 of grad_input in subsequent computations. grad_input will only correspond to the inputs given as positional arguments and all kwarg arguments are ignored. Entries in grad_input and grad_output will be None 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 existing backward hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing backward hooks on this torch.nn.modules.Module. Note that global backward hooks registered with register_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, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle

Register 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 of grad_output in subsequent computations. Entries in grad_output will be None 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 existing backward_pre hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing backward_pre hooks on this torch.nn.modules.Module. Note that global backward_pre hooks registered with register_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)

Register 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 the incompatible_keys argument is a NamedTuple consisting of attributes missing_keys and unexpected_keys. missing_keys is a list of str containing the missing keys and unexpected_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 load_state_dict() with strict=True are affected by modifications the hook makes to missing_keys or unexpected_keys, as expected. Additions to either set of keys will result in an error being thrown when strict=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_load_state_dict_pre_hook(hook)

Register a pre-hook to be run before module’s load_state_dict() is called.

It should have the following signature::

hook(module, state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs) -> None # noqa: B950

Parameters:

hook (Callable) – Callable hook that will be invoked before loading the state dict.

register_module(name: str, module: Module | None) → None

Alias for add_module().

register_parameter(name: str, param: Parameter | None) → None

Add a parameter to the module.

The parameter can be accessed as an attribute using given name.

Parameters:

• name (str) – name of the parameter. The parameter can be accessed from this module using the given name

• param (Parameter or None) – parameter to be added to the module. If None, then operations that run on parameters, such as cuda, are ignored. If None, the parameter is not included in the module’s state_dict.

register_state_dict_post_hook(hook)

Register a post-hook for the state_dict() method.

It should have the following signature::

hook(module, state_dict, prefix, local_metadata) -> None

The registered hooks can modify the state_dict inplace.

register_state_dict_pre_hook(hook)

Register a pre-hook for the state_dict() method.

It should have the following signature::

hook(module, prefix, keep_vars) -> None

The registered hooks can be used to perform pre-processing before the state_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-disable-grad-doc for a comparison between .requires_grad_() and several similar mechanisms that may be confused with it.

Parameters:

requires_grad (bool) – whether autograd should record operations on parameters in this module. Default: True.

Returns:

self

Return type:

Module

set_extra_state(state: Any) → None

Set extra state contained in the loaded state_dict.

This function is called from load_state_dict() to handle any extra state found within the state_dict. Implement this function and a corresponding get_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

set_submodule(target: str, module: Module) → None

Set the submodule given by target if it exists, otherwise throw 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 submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.)

To overide the Conv2d with a new submodule Linear, you would call set_submodule("net_b.net_c.conv", nn.Linear(33, 16)).

Parameters:

• target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.)

• module – The module to set the submodule to.

Raises:

• ValueError – If the target string is empty

• AttributeError – If the target string references an invalid path or resolves to something that is not an nn.Module

share_memory() → T

See torch.Tensor.share_memory_().

state_dict(*args, destination=None, prefix='', keep_vars=False)

Return a dictionary containing references to the whole state of the module.

Both parameters and persistent buffers (e.g. running averages) are included. 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 references to the module’s parameters and buffers.

Warning

Currently state_dict() also accepts positional arguments for destination, prefix and keep_vars in order. However, this is being deprecated and keyword arguments will be enforced in future releases.

Warning

Please avoid the use of argument destination as it is not designed for end-users.

Parameters:

• destination (dict, optional) – If provided, the state of module will be 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 buffer names to compose the keys in state_dict. Default: ''.

• keep_vars (bool, optional) – by default the Tensor s returned in the state dict are detached from autograd. If it’s set to True, detaching will not be performed. Default: False.

Returns:

a dictionary containing a whole state of the module

Return type:

dict

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> module.state_dict().keys()
['bias', 'weight']

to(*args, **kwargs)

Move and/or cast 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 complex dtypes. In addition, this method will only cast the floating point or complex parameters and buffers to dtype (if given). The integral parameters and buffers will be moved device, if that is given, but with dtypes unchanged. When non_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 module

• dtype (torch.dtype) – the desired floating point or complex dtype of the parameters and buffers in this module

• tensor (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:

Module

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: int | str | device | None, recurse: bool = True) → T

Move 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:

Module

train(mode: bool = True) → T

Set the module in training 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.

Parameters:

mode (bool) – whether to set training mode (True) or evaluation mode (False). Default: True.

Returns:

self

Return type:

Module

training: bool

type(dst_type: dtype | str) → T

Casts all parameters and buffers to dst_type.

Note

This method modifies the module in-place.

Parameters:

dst_type (type or string) – the desired type

Returns:

self

Return type:

Module

xpu(device: int | device | None = None) → T

Move 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.

Parameters:

device (int, optional) – if specified, all parameters will be copied to that device

Returns:

self

Return type:

Module

zero_grad(set_to_none: bool = True) → None

Reset 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.

class lzero.model.stochastic_muzero_model.ChanceEncoderBackboneMLP(input_dimensions, chance_encoding_dim=4)

Bases: Module

Overview:

The definition of chance encoder backbone network, which is used to encode the (vector) observation into a latent space.

Parameters:

• input_dimensions (-) – The dimension of observation space.

• chance_encoding_dim (-) – The dimension of chance encoding.

T_destination = ~T_destination

__init__(input_dimensions, chance_encoding_dim=4)

Initialize internal Module state, shared by both nn.Module and ScriptModule.

_apply(fn, recurse=True)

_backward_hooks: Dict[int, Callable]

_backward_pre_hooks: Dict[int, Callable]

_buffers: Dict[str, Tensor | None]

_call_impl(*args, **kwargs)

_compiled_call_impl: Callable | None = None

_forward_hooks: Dict[int, Callable]

_forward_hooks_always_called: Dict[int, bool]

_forward_hooks_with_kwargs: Dict[int, bool]

_forward_pre_hooks: Dict[int, Callable]

_forward_pre_hooks_with_kwargs: Dict[int, bool]

_get_backward_hooks()

Return the backward hooks for use in the call function.

It returns two lists, one with the full backward hooks and one with the non-full backward hooks.

_get_backward_pre_hooks()

_get_name()

_is_full_backward_hook: bool | None

_load_from_state_dict(state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs)

Copy parameters and buffers from state_dict into only this module, but not its descendants.

This is called on every submodule in load_state_dict(). Metadata saved for this module in input state_dict is provided as local_metadata. For state dicts without metadata, local_metadata is empty. Subclasses can achieve class-specific backward compatible loading using the version number at local_metadata.get(“version”, None). Additionally, local_metadata can also contain the key assign_to_params_buffers that indicates whether keys should be assigned their corresponding tensor in the state_dict.

Note

state_dict is not the same object as the input state_dict to load_state_dict(). So it can be modified.

Parameters:

• state_dict (dict) – a dict containing parameters and persistent buffers.

• prefix (str) – the prefix for parameters and buffers used in this module

• local_metadata (dict) – a dict containing the metadata for this module. See

• strict (bool) – whether to strictly enforce that the keys in state_dict with prefix match the names of parameters and buffers in this module

• missing_keys (list of str) – if strict=True, add missing keys to this list

• unexpected_keys (list of str) – if strict=True, add unexpected keys to this list

• error_msgs (list of str) – error messages should be added to this list, and will be reported together in load_state_dict()

_load_state_dict_post_hooks: Dict[int, Callable]

_load_state_dict_pre_hooks: Dict[int, Callable]

_maybe_warn_non_full_backward_hook(inputs, result, grad_fn)

_modules: Dict[str, 'Module' | None]

_named_members(get_members_fn, prefix='', recurse=True, remove_duplicate: bool = True)

Help yield various names + members of modules.

_non_persistent_buffers_set: Set[str]

_parameters: Dict[str, Parameter | None]

_register_load_state_dict_pre_hook(hook, with_module=False)

See register_load_state_dict_pre_hook() for details.

A subtle difference is that if with_module is set to False, then the hook will not take the module as the first argument whereas register_load_state_dict_pre_hook() always takes the module as the first argument.

Parameters:

• hook (Callable) – Callable hook that will be invoked before loading the state dict.

• with_module (bool, optional) – Whether or not to pass the module instance to the hook as the first parameter.

_register_state_dict_hook(hook)

Register a post-hook for the state_dict() method.

It should have the following signature::

hook(module, state_dict, prefix, local_metadata) -> None or state_dict

The registered hooks can modify the state_dict inplace or return a new one. If a new state_dict is returned, it will only be respected if it is the root module that state_dict() is called from.

_replicate_for_data_parallel()

_save_to_state_dict(destination, prefix, keep_vars)

Save module state to the destination dictionary.

The destination dictionary will contain the state of the module, but not its descendants. This is called on every submodule in state_dict().

In rare cases, subclasses can achieve class-specific behavior by overriding this method with custom logic.

Parameters:

• destination (dict) – a dict where state will be stored

• prefix (str) – the prefix for parameters and buffers used in this module

_slow_forward(*input, **kwargs)

_state_dict_hooks: Dict[int, Callable]

_state_dict_pre_hooks: Dict[int, Callable]

_version: int = 1

This allows better BC support for load_state_dict(). In state_dict(), the version number will be saved as in the attribute _metadata of the returned state dict, and thus pickled. _metadata is a dictionary 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 shall be bumped, and the module’s _load_from_state_dict method can compare the version number and do appropriate changes if the state dict is from before the change.

_wrapped_call_impl(*args, **kwargs)

add_module(name: str, module: Module | None) → None

Add a child module to the current module.

The module can be accessed as an attribute using the given name.

Parameters:

• name (str) – name of the child module. The child module can be accessed from this module using the given name

• module (Module) – child module to be added to the module.

apply(fn: Callable[[Module], None]) → T

Apply fn recursively to every submodule (as returned by .children()) as well as self.

Typical use includes initializing the parameters of a model (see also nn-init-doc).

Parameters:

fn (Module -> None) – function to be applied to each submodule

Returns:

self

Return type:

Module

Example:

>>> @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_features=2, out_features=2, bias=True)
Parameter containing:
tensor([[1., 1.],
        [1., 1.]], requires_grad=True)
Linear(in_features=2, out_features=2, bias=True)
Parameter containing:
tensor([[1., 1.],
        [1., 1.]], requires_grad=True)
Sequential(
  (0): Linear(in_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
)

bfloat16() → T

Casts all floating point parameters and buffers to bfloat16 datatype.

Note

This method modifies the module in-place.

Returns:

self

Return type:

Module

buffers(recurse: bool = True) → Iterator[Tensor]

Return an iterator over module buffers.

Parameters:

recurse (bool) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module.

Yields:

torch.Tensor – module buffer

Example:

>>> # 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)

call_super_init: bool = False

children() → Iterator[Module]

Return 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

Move all model parameters and buffers to the CPU.

Note

This method modifies the module in-place.

Returns:

self

Return type:

Module

cuda(device: int | device | None = None) → T

Move 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.

Parameters:

device (int, optional) – if specified, all parameters will be copied to that device

Returns:

self

Return type:

Module

double() → T

Casts all floating point parameters and buffers to double datatype.

Note

This method modifies the module in-place.

Returns:

self

Return type:

Module

dump_patches: bool = False

eval() → T

Set 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-disable-grad-doc for a comparison between .eval() and several similar mechanisms that may be confused with it.

Returns:

self

Return type:

Module

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:

Module

forward(x)

Define the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

get_buffer(target: str) → Tensor

Return the buffer given by target if it exists, otherwise throw an error.

See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target.

Parameters:

target – The fully-qualified string name of the buffer to look for. (See get_submodule for how to specify a fully-qualified string.)

Returns:

The buffer referenced by target

Return type:

torch.Tensor

Raises:

AttributeError – If the target string references an invalid path or resolves to something that is not a buffer

get_extra_state() → Any

Return 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:

object

get_parameter(target: str) → Parameter

Return the parameter given by target if it exists, otherwise throw an error.

See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target.

Parameters:

target – The fully-qualified string name of the Parameter to look for. (See get_submodule for how to specify a fully-qualified string.)

Returns:

The Parameter referenced by target

Return type:

torch.nn.Parameter

Raises:

AttributeError – If the target string references an invalid path or resolves to something that is not an nn.Parameter

get_submodule(target: str) → Module

Return the submodule given by target if it exists, otherwise throw 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 submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.)

To check whether or not we have the linear submodule, we would call get_submodule("net_b.linear"). To check whether we have the conv submodule, we would call get_submodule("net_b.net_c.conv").

The runtime of get_submodule is bounded by the degree of module nesting in target. A query against named_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:

torch.nn.Module

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:

Module

ipu(device: int | device | None = None) → T

Move 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.

Parameters:

device (int, optional) – if specified, all parameters will be copied to that device

Returns:

self

Return type:

Module

load_state_dict(state_dict: Mapping[str, Any], strict: bool = True, assign: bool = False)

Copy parameters and buffers from state_dict into this module and its descendants.

If strict is True, then the keys of state_dict must exactly match the keys returned by this module’s state_dict() function.

Warning

If assign is True the optimizer must be created after the call to load_state_dict unless get_swap_module_params_on_conversion() is True.

Parameters:

• state_dict (dict) – a dict containing parameters and persistent buffers.

• strict (bool, optional) – whether to strictly enforce that the keys in state_dict match the keys returned by this module’s state_dict() function. Default: True

• assign (bool, optional) – When False, the properties of the tensors in the current module are preserved while when True, the properties of the Tensors in the state dict are preserved. The only exception is the requires_grad field of Default: ``False`

Returns:

• missing_keys is a list of str containing any keys that are expected

  by this module but missing from the provided state_dict.

• unexpected_keys is a list of str containing the keys that are not

  expected by this module but present in the provided state_dict.

Return type:

NamedTuple with missing_keys and unexpected_keys fields

Note

If a parameter or buffer is registered as None and its corresponding key exists in state_dict, load_state_dict() will raise a RuntimeError.

modules() → Iterator[Module]

Return an iterator over all modules in the network.

Yields:

Module – a module in the network

Note

Duplicate modules are returned only once. In the following example, 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_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
)
1 -> Linear(in_features=2, out_features=2, bias=True)

mtia(device: int | device | None = None) → T

Move all model parameters and buffers to the MTIA.

This also makes associated parameters and buffers different objects. So it should be called before constructing optimizer if the module will live on MTIA while being optimized.

Note

This method modifies the module in-place.

Parameters:

device (int, optional) – if specified, all parameters will be copied to that device

Returns:

self

Return type:

Module

named_buffers(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[Tuple[str, Tensor]]

Return an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.

Parameters:

• prefix (str) – prefix to prepend to all buffer names.

• recurse (bool, optional) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are 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 buffer

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, buf in self.named_buffers():
>>>     if name in ['running_var']:
>>>         print(buf.size())

named_children() → Iterator[Tuple[str, Module]]

Return an iterator over immediate children modules, yielding both the name of the module as well as the module itself.

Yields:

(str, Module) – Tuple containing a name and child module

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, module in model.named_children():
>>>     if name in ['conv4', 'conv5']:
>>>         print(module)

named_modules(memo: Set[Module] | None = None, prefix: str = '', remove_duplicate: bool = True)

Return an iterator over all modules in the network, yielding both the name of the module as well as the module itself.

Parameters:

• memo – a memo to store the set of modules already added to the result

• prefix – a prefix that will be added to the name of the module

• remove_duplicate – whether to remove the duplicated module instances in the result or not

Yields:

(str, Module) – Tuple of name and module

Note

Duplicate modules are returned only once. In the following example, 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_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
))
1 -> ('0', Linear(in_features=2, out_features=2, bias=True))

named_parameters(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[Tuple[str, Parameter]]

Return an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.

Parameters:

• prefix (str) – prefix to prepend to all parameter names.

• recurse (bool) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module.

• remove_duplicate (bool, optional) – whether to remove the duplicated parameters in the result. Defaults to True.

Yields:

(str, Parameter) – Tuple containing the name and parameter

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, param in self.named_parameters():
>>>     if name in ['bias']:
>>>         print(param.size())

parameters(recurse: bool = True) → Iterator[Parameter]

Return an iterator over module parameters.

This is typically passed to an optimizer.

Parameters:

recurse (bool) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module.

Yields:

Parameter – module parameter

Example:

>>> # 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)

register_backward_hook(hook: Callable[[Module, Tuple[Tensor, ...] | Tensor, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor]) → RemovableHandle

Register 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_buffer(name: str, tensor: Tensor | None, persistent: bool = True) → None

Add a buffer to the module.

This is typically used to register a buffer that should not to be considered a model parameter. For example, BatchNorm’s running_mean is not a parameter, but is part of the module’s state. Buffers, by default, are persistent and will be saved alongside parameters. This behavior can be changed by setting persistent to False. The only difference between a persistent buffer and a non-persistent buffer is that the latter will not be a part of this module’s state_dict.

Buffers can be accessed as attributes using given names.

Parameters:

• name (str) – name of the buffer. The buffer can be accessed from this module using the given name

• tensor (Tensor or None) – buffer to be registered. If None, then operations that run on buffers, such as cuda, are ignored. If None, the buffer is not included in the module’s state_dict.

• persistent (bool) – whether the buffer is part of this module’s state_dict.

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> self.register_buffer('running_mean', torch.zeros(num_features))

register_forward_hook(hook: Callable[[T, Tuple[Any, ...], Any], Any | None] | Callable[[T, Tuple[Any, ...], Dict[str, Any], Any], Any | None], *, prepend: bool = False, with_kwargs: bool = False, always_call: bool = False) → RemovableHandle

Register a forward hook on the module.

The hook will be called every time after forward() has computed an output.

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 the forward. The hook can modify the output. It can modify the input inplace but it will not have effect on forward since this is called after forward() is called. The hook should have the following signature:

hook(module, args, output) -> None or modified output

If with_kwargs is True, the forward hook will be passed the kwargs 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 provided hook will be fired before all existing forward hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing forward hooks on this torch.nn.modules.Module. Note that global forward hooks registered with register_module_forward_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

• always_call (bool) – If True the hook 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: Callable[[T, Tuple[Any, ...]], Any | None] | Callable[[T, Tuple[Any, ...], Dict[str, Any]], Tuple[Any, Dict[str, Any]] | None], *, prepend: bool = False, with_kwargs: bool = False) → RemovableHandle

Register 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 the forward. 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 existing forward_pre hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing forward_pre hooks on this torch.nn.modules.Module. Note that global forward_pre hooks registered with register_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, Tuple[Tensor, ...] | Tensor, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle

Register 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 and grad_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 of grad_input in subsequent computations. grad_input will only correspond to the inputs given as positional arguments and all kwarg arguments are ignored. Entries in grad_input and grad_output will be None 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 existing backward hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing backward hooks on this torch.nn.modules.Module. Note that global backward hooks registered with register_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, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle

Register 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 of grad_output in subsequent computations. Entries in grad_output will be None 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 existing backward_pre hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing backward_pre hooks on this torch.nn.modules.Module. Note that global backward_pre hooks registered with register_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)

Register 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 the incompatible_keys argument is a NamedTuple consisting of attributes missing_keys and unexpected_keys. missing_keys is a list of str containing the missing keys and unexpected_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 load_state_dict() with strict=True are affected by modifications the hook makes to missing_keys or unexpected_keys, as expected. Additions to either set of keys will result in an error being thrown when strict=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_load_state_dict_pre_hook(hook)

Register a pre-hook to be run before module’s load_state_dict() is called.

It should have the following signature::

hook(module, state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs) -> None # noqa: B950

Parameters:

hook (Callable) – Callable hook that will be invoked before loading the state dict.

register_module(name: str, module: Module | None) → None

Alias for add_module().

register_parameter(name: str, param: Parameter | None) → None

Add a parameter to the module.

The parameter can be accessed as an attribute using given name.

Parameters:

• name (str) – name of the parameter. The parameter can be accessed from this module using the given name

• param (Parameter or None) – parameter to be added to the module. If None, then operations that run on parameters, such as cuda, are ignored. If None, the parameter is not included in the module’s state_dict.

register_state_dict_post_hook(hook)

Register a post-hook for the state_dict() method.

It should have the following signature::

hook(module, state_dict, prefix, local_metadata) -> None

The registered hooks can modify the state_dict inplace.

register_state_dict_pre_hook(hook)

Register a pre-hook for the state_dict() method.

It should have the following signature::

hook(module, prefix, keep_vars) -> None

The registered hooks can be used to perform pre-processing before the state_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-disable-grad-doc for a comparison between .requires_grad_() and several similar mechanisms that may be confused with it.

Parameters:

requires_grad (bool) – whether autograd should record operations on parameters in this module. Default: True.

Returns:

self

Return type:

Module

set_extra_state(state: Any) → None

Set extra state contained in the loaded state_dict.

This function is called from load_state_dict() to handle any extra state found within the state_dict. Implement this function and a corresponding get_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

set_submodule(target: str, module: Module) → None

Set the submodule given by target if it exists, otherwise throw 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 submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.)

To overide the Conv2d with a new submodule Linear, you would call set_submodule("net_b.net_c.conv", nn.Linear(33, 16)).

Parameters:

• target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.)

• module – The module to set the submodule to.

Raises:

• ValueError – If the target string is empty

• AttributeError – If the target string references an invalid path or resolves to something that is not an nn.Module

share_memory() → T

See torch.Tensor.share_memory_().

state_dict(*args, destination=None, prefix='', keep_vars=False)

Return a dictionary containing references to the whole state of the module.

Both parameters and persistent buffers (e.g. running averages) are included. 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 references to the module’s parameters and buffers.

Warning

Currently state_dict() also accepts positional arguments for destination, prefix and keep_vars in order. However, this is being deprecated and keyword arguments will be enforced in future releases.

Warning

Please avoid the use of argument destination as it is not designed for end-users.

Parameters:

• destination (dict, optional) – If provided, the state of module will be 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 buffer names to compose the keys in state_dict. Default: ''.

• keep_vars (bool, optional) – by default the Tensor s returned in the state dict are detached from autograd. If it’s set to True, detaching will not be performed. Default: False.

Returns:

a dictionary containing a whole state of the module

Return type:

dict

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> module.state_dict().keys()
['bias', 'weight']

to(*args, **kwargs)

Move and/or cast 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 complex dtypes. In addition, this method will only cast the floating point or complex parameters and buffers to dtype (if given). The integral parameters and buffers will be moved device, if that is given, but with dtypes unchanged. When non_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 module

• dtype (torch.dtype) – the desired floating point or complex dtype of the parameters and buffers in this module

• tensor (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:

Module

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: int | str | device | None, recurse: bool = True) → T

Move 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:

Module

train(mode: bool = True) → T

Set the module in training 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.

Parameters:

mode (bool) – whether to set training mode (True) or evaluation mode (False). Default: True.

Returns:

self

Return type:

Module

training: bool

type(dst_type: dtype | str) → T

Casts all parameters and buffers to dst_type.

Note

This method modifies the module in-place.

Parameters:

dst_type (type or string) – the desired type

Returns:

self

Return type:

Module

xpu(device: int | device | None = None) → T

Move 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.

Parameters:

device (int, optional) – if specified, all parameters will be copied to that device

Returns:

self

Return type:

Module

zero_grad(set_to_none: bool = True) → None

Reset 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.

class lzero.model.stochastic_muzero_model.ChanceEncoder(input_dimensions, action_dimension, encoder_backbone_type='conv')

Bases: Module

T_destination = ~T_destination

__init__(input_dimensions, action_dimension, encoder_backbone_type='conv')

Initialize internal Module state, shared by both nn.Module and ScriptModule.

_apply(fn, recurse=True)

_backward_hooks: Dict[int, Callable]

_backward_pre_hooks: Dict[int, Callable]

_buffers: Dict[str, Tensor | None]

_call_impl(*args, **kwargs)

_compiled_call_impl: Callable | None = None

_forward_hooks: Dict[int, Callable]

_forward_hooks_always_called: Dict[int, bool]

_forward_hooks_with_kwargs: Dict[int, bool]

_forward_pre_hooks: Dict[int, Callable]

_forward_pre_hooks_with_kwargs: Dict[int, bool]

_get_backward_hooks()

Return the backward hooks for use in the call function.

It returns two lists, one with the full backward hooks and one with the non-full backward hooks.

_get_backward_pre_hooks()

_get_name()

_is_full_backward_hook: bool | None

_load_from_state_dict(state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs)

Copy parameters and buffers from state_dict into only this module, but not its descendants.

This is called on every submodule in load_state_dict(). Metadata saved for this module in input state_dict is provided as local_metadata. For state dicts without metadata, local_metadata is empty. Subclasses can achieve class-specific backward compatible loading using the version number at local_metadata.get(“version”, None). Additionally, local_metadata can also contain the key assign_to_params_buffers that indicates whether keys should be assigned their corresponding tensor in the state_dict.

Note

state_dict is not the same object as the input state_dict to load_state_dict(). So it can be modified.

Parameters:

• state_dict (dict) – a dict containing parameters and persistent buffers.

• prefix (str) – the prefix for parameters and buffers used in this module

• local_metadata (dict) – a dict containing the metadata for this module. See

• strict (bool) – whether to strictly enforce that the keys in state_dict with prefix match the names of parameters and buffers in this module

• missing_keys (list of str) – if strict=True, add missing keys to this list

• unexpected_keys (list of str) – if strict=True, add unexpected keys to this list

• error_msgs (list of str) – error messages should be added to this list, and will be reported together in load_state_dict()

_load_state_dict_post_hooks: Dict[int, Callable]

_load_state_dict_pre_hooks: Dict[int, Callable]

_maybe_warn_non_full_backward_hook(inputs, result, grad_fn)

_modules: Dict[str, 'Module' | None]

_named_members(get_members_fn, prefix='', recurse=True, remove_duplicate: bool = True)

Help yield various names + members of modules.

_non_persistent_buffers_set: Set[str]

_parameters: Dict[str, Parameter | None]

_register_load_state_dict_pre_hook(hook, with_module=False)

See register_load_state_dict_pre_hook() for details.

A subtle difference is that if with_module is set to False, then the hook will not take the module as the first argument whereas register_load_state_dict_pre_hook() always takes the module as the first argument.

Parameters:

• hook (Callable) – Callable hook that will be invoked before loading the state dict.

• with_module (bool, optional) – Whether or not to pass the module instance to the hook as the first parameter.

_register_state_dict_hook(hook)

Register a post-hook for the state_dict() method.

It should have the following signature::

hook(module, state_dict, prefix, local_metadata) -> None or state_dict

The registered hooks can modify the state_dict inplace or return a new one. If a new state_dict is returned, it will only be respected if it is the root module that state_dict() is called from.

_replicate_for_data_parallel()

_save_to_state_dict(destination, prefix, keep_vars)

Save module state to the destination dictionary.

The destination dictionary will contain the state of the module, but not its descendants. This is called on every submodule in state_dict().

In rare cases, subclasses can achieve class-specific behavior by overriding this method with custom logic.

Parameters:

• destination (dict) – a dict where state will be stored

• prefix (str) – the prefix for parameters and buffers used in this module

_slow_forward(*input, **kwargs)

_state_dict_hooks: Dict[int, Callable]

_state_dict_pre_hooks: Dict[int, Callable]

_version: int = 1

This allows better BC support for load_state_dict(). In state_dict(), the version number will be saved as in the attribute _metadata of the returned state dict, and thus pickled. _metadata is a dictionary 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 shall be bumped, and the module’s _load_from_state_dict method can compare the version number and do appropriate changes if the state dict is from before the change.

_wrapped_call_impl(*args, **kwargs)

add_module(name: str, module: Module | None) → None

Add a child module to the current module.

The module can be accessed as an attribute using the given name.

Parameters:

• name (str) – name of the child module. The child module can be accessed from this module using the given name

• module (Module) – child module to be added to the module.

apply(fn: Callable[[Module], None]) → T

Apply fn recursively to every submodule (as returned by .children()) as well as self.

Typical use includes initializing the parameters of a model (see also nn-init-doc).

Parameters:

fn (Module -> None) – function to be applied to each submodule

Returns:

self

Return type:

Module

Example:

>>> @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_features=2, out_features=2, bias=True)
Parameter containing:
tensor([[1., 1.],
        [1., 1.]], requires_grad=True)
Linear(in_features=2, out_features=2, bias=True)
Parameter containing:
tensor([[1., 1.],
        [1., 1.]], requires_grad=True)
Sequential(
  (0): Linear(in_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
)

bfloat16() → T

Casts all floating point parameters and buffers to bfloat16 datatype.

Note

This method modifies the module in-place.

Returns:

self

Return type:

Module

buffers(recurse: bool = True) → Iterator[Tensor]

Return an iterator over module buffers.

Parameters:

recurse (bool) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module.

Yields:

torch.Tensor – module buffer

Example:

>>> # 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)

call_super_init: bool = False

children() → Iterator[Module]

Return 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

Move all model parameters and buffers to the CPU.

Note

This method modifies the module in-place.

Returns:

self

Return type:

Module

cuda(device: int | device | None = None) → T

Move 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.

Parameters:

device (int, optional) – if specified, all parameters will be copied to that device

Returns:

self

Return type:

Module

double() → T

Casts all floating point parameters and buffers to double datatype.

Note

This method modifies the module in-place.

Returns:

self

Return type:

Module

dump_patches: bool = False

eval() → T

Set 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-disable-grad-doc for a comparison between .eval() and several similar mechanisms that may be confused with it.

Returns:

self

Return type:

Module

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:

Module

forward(observations)

Overview:

Forward method for the ChanceEncoder. This method takes an observation and applies the encoder to transform it to a latent space. Then applies the StraightThroughEstimator to this encoding. References: Planning in Stochastic Environments with a Learned Model (ICLR 2022), page 5, Chance Outcomes section.

Parameters:

observations (-) – Observation tensor.

Returns:

Transformed tensor after applying one-hot argmax. - chance_encoding (torch.Tensor): Encoding of the input observation tensor.

Return type:

• chance (torch.Tensor)

get_buffer(target: str) → Tensor

Return the buffer given by target if it exists, otherwise throw an error.

See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target.

Parameters:

target – The fully-qualified string name of the buffer to look for. (See get_submodule for how to specify a fully-qualified string.)

Returns:

The buffer referenced by target

Return type:

torch.Tensor

Raises:

AttributeError – If the target string references an invalid path or resolves to something that is not a buffer

get_extra_state() → Any

Return 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:

object

get_parameter(target: str) → Parameter

Return the parameter given by target if it exists, otherwise throw an error.

See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target.

Parameters:

target – The fully-qualified string name of the Parameter to look for. (See get_submodule for how to specify a fully-qualified string.)

Returns:

The Parameter referenced by target

Return type:

torch.nn.Parameter

Raises:

AttributeError – If the target string references an invalid path or resolves to something that is not an nn.Parameter

get_submodule(target: str) → Module

Return the submodule given by target if it exists, otherwise throw 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 submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.)

To check whether or not we have the linear submodule, we would call get_submodule("net_b.linear"). To check whether we have the conv submodule, we would call get_submodule("net_b.net_c.conv").

The runtime of get_submodule is bounded by the degree of module nesting in target. A query against named_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:

torch.nn.Module

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:

Module

ipu(device: int | device | None = None) → T

Move 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.

Parameters:

device (int, optional) – if specified, all parameters will be copied to that device

Returns:

self

Return type:

Module

load_state_dict(state_dict: Mapping[str, Any], strict: bool = True, assign: bool = False)

Copy parameters and buffers from state_dict into this module and its descendants.

If strict is True, then the keys of state_dict must exactly match the keys returned by this module’s state_dict() function.

Warning

If assign is True the optimizer must be created after the call to load_state_dict unless get_swap_module_params_on_conversion() is True.

Parameters:

• state_dict (dict) – a dict containing parameters and persistent buffers.

• strict (bool, optional) – whether to strictly enforce that the keys in state_dict match the keys returned by this module’s state_dict() function. Default: True

• assign (bool, optional) – When False, the properties of the tensors in the current module are preserved while when True, the properties of the Tensors in the state dict are preserved. The only exception is the requires_grad field of Default: ``False`

Returns:

• missing_keys is a list of str containing any keys that are expected

  by this module but missing from the provided state_dict.

• unexpected_keys is a list of str containing the keys that are not

  expected by this module but present in the provided state_dict.

Return type:

NamedTuple with missing_keys and unexpected_keys fields

Note

If a parameter or buffer is registered as None and its corresponding key exists in state_dict, load_state_dict() will raise a RuntimeError.

modules() → Iterator[Module]

Return an iterator over all modules in the network.

Yields:

Module – a module in the network

Note

Duplicate modules are returned only once. In the following example, 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_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
)
1 -> Linear(in_features=2, out_features=2, bias=True)

mtia(device: int | device | None = None) → T

Move all model parameters and buffers to the MTIA.

This also makes associated parameters and buffers different objects. So it should be called before constructing optimizer if the module will live on MTIA while being optimized.

Note

This method modifies the module in-place.

Parameters:

device (int, optional) – if specified, all parameters will be copied to that device

Returns:

self

Return type:

Module

named_buffers(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[Tuple[str, Tensor]]

Return an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.

Parameters:

• prefix (str) – prefix to prepend to all buffer names.

• recurse (bool, optional) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are 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 buffer

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, buf in self.named_buffers():
>>>     if name in ['running_var']:
>>>         print(buf.size())

named_children() → Iterator[Tuple[str, Module]]

Return an iterator over immediate children modules, yielding both the name of the module as well as the module itself.

Yields:

(str, Module) – Tuple containing a name and child module

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, module in model.named_children():
>>>     if name in ['conv4', 'conv5']:
>>>         print(module)

named_modules(memo: Set[Module] | None = None, prefix: str = '', remove_duplicate: bool = True)

Return an iterator over all modules in the network, yielding both the name of the module as well as the module itself.

Parameters:

• memo – a memo to store the set of modules already added to the result

• prefix – a prefix that will be added to the name of the module

• remove_duplicate – whether to remove the duplicated module instances in the result or not

Yields:

(str, Module) – Tuple of name and module

Note

Duplicate modules are returned only once. In the following example, 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_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
))
1 -> ('0', Linear(in_features=2, out_features=2, bias=True))

named_parameters(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[Tuple[str, Parameter]]

Return an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.

Parameters:

• prefix (str) – prefix to prepend to all parameter names.

• recurse (bool) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module.

• remove_duplicate (bool, optional) – whether to remove the duplicated parameters in the result. Defaults to True.

Yields:

(str, Parameter) – Tuple containing the name and parameter

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, param in self.named_parameters():
>>>     if name in ['bias']:
>>>         print(param.size())

parameters(recurse: bool = True) → Iterator[Parameter]

Return an iterator over module parameters.

This is typically passed to an optimizer.

Parameters:

recurse (bool) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module.

Yields:

Parameter – module parameter

Example:

>>> # 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)

register_backward_hook(hook: Callable[[Module, Tuple[Tensor, ...] | Tensor, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor]) → RemovableHandle

Register 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_buffer(name: str, tensor: Tensor | None, persistent: bool = True) → None

Add a buffer to the module.

This is typically used to register a buffer that should not to be considered a model parameter. For example, BatchNorm’s running_mean is not a parameter, but is part of the module’s state. Buffers, by default, are persistent and will be saved alongside parameters. This behavior can be changed by setting persistent to False. The only difference between a persistent buffer and a non-persistent buffer is that the latter will not be a part of this module’s state_dict.

Buffers can be accessed as attributes using given names.

Parameters:

• name (str) – name of the buffer. The buffer can be accessed from this module using the given name

• tensor (Tensor or None) – buffer to be registered. If None, then operations that run on buffers, such as cuda, are ignored. If None, the buffer is not included in the module’s state_dict.

• persistent (bool) – whether the buffer is part of this module’s state_dict.

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> self.register_buffer('running_mean', torch.zeros(num_features))

register_forward_hook(hook: Callable[[T, Tuple[Any, ...], Any], Any | None] | Callable[[T, Tuple[Any, ...], Dict[str, Any], Any], Any | None], *, prepend: bool = False, with_kwargs: bool = False, always_call: bool = False) → RemovableHandle

Register a forward hook on the module.

The hook will be called every time after forward() has computed an output.

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 the forward. The hook can modify the output. It can modify the input inplace but it will not have effect on forward since this is called after forward() is called. The hook should have the following signature:

hook(module, args, output) -> None or modified output

If with_kwargs is True, the forward hook will be passed the kwargs 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 provided hook will be fired before all existing forward hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing forward hooks on this torch.nn.modules.Module. Note that global forward hooks registered with register_module_forward_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

• always_call (bool) – If True the hook 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: Callable[[T, Tuple[Any, ...]], Any | None] | Callable[[T, Tuple[Any, ...], Dict[str, Any]], Tuple[Any, Dict[str, Any]] | None], *, prepend: bool = False, with_kwargs: bool = False) → RemovableHandle

Register 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 the forward. 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 existing forward_pre hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing forward_pre hooks on this torch.nn.modules.Module. Note that global forward_pre hooks registered with register_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, Tuple[Tensor, ...] | Tensor, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle

Register 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 and grad_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 of grad_input in subsequent computations. grad_input will only correspond to the inputs given as positional arguments and all kwarg arguments are ignored. Entries in grad_input and grad_output will be None 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 existing backward hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing backward hooks on this torch.nn.modules.Module. Note that global backward hooks registered with register_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, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle

Register 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 of grad_output in subsequent computations. Entries in grad_output will be None 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 existing backward_pre hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing backward_pre hooks on this torch.nn.modules.Module. Note that global backward_pre hooks registered with register_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)

Register 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 the incompatible_keys argument is a NamedTuple consisting of attributes missing_keys and unexpected_keys. missing_keys is a list of str containing the missing keys and unexpected_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 load_state_dict() with strict=True are affected by modifications the hook makes to missing_keys or unexpected_keys, as expected. Additions to either set of keys will result in an error being thrown when strict=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_load_state_dict_pre_hook(hook)

Register a pre-hook to be run before module’s load_state_dict() is called.

It should have the following signature::

hook(module, state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs) -> None # noqa: B950

Parameters:

hook (Callable) – Callable hook that will be invoked before loading the state dict.

register_module(name: str, module: Module | None) → None

Alias for add_module().

register_parameter(name: str, param: Parameter | None) → None

Add a parameter to the module.

The parameter can be accessed as an attribute using given name.

Parameters:

• name (str) – name of the parameter. The parameter can be accessed from this module using the given name

• param (Parameter or None) – parameter to be added to the module. If None, then operations that run on parameters, such as cuda, are ignored. If None, the parameter is not included in the module’s state_dict.

register_state_dict_post_hook(hook)

Register a post-hook for the state_dict() method.

It should have the following signature::

hook(module, state_dict, prefix, local_metadata) -> None

The registered hooks can modify the state_dict inplace.

register_state_dict_pre_hook(hook)

Register a pre-hook for the state_dict() method.

It should have the following signature::

hook(module, prefix, keep_vars) -> None

The registered hooks can be used to perform pre-processing before the state_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-disable-grad-doc for a comparison between .requires_grad_() and several similar mechanisms that may be confused with it.

Parameters:

requires_grad (bool) – whether autograd should record operations on parameters in this module. Default: True.

Returns:

self

Return type:

Module

set_extra_state(state: Any) → None

Set extra state contained in the loaded state_dict.

This function is called from load_state_dict() to handle any extra state found within the state_dict. Implement this function and a corresponding get_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

set_submodule(target: str, module: Module) → None

Set the submodule given by target if it exists, otherwise throw 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 submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.)

To overide the Conv2d with a new submodule Linear, you would call set_submodule("net_b.net_c.conv", nn.Linear(33, 16)).

Parameters:

• target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.)

• module – The module to set the submodule to.

Raises:

• ValueError – If the target string is empty

• AttributeError – If the target string references an invalid path or resolves to something that is not an nn.Module

share_memory() → T

See torch.Tensor.share_memory_().

state_dict(*args, destination=None, prefix='', keep_vars=False)

Return a dictionary containing references to the whole state of the module.

Both parameters and persistent buffers (e.g. running averages) are included. 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 references to the module’s parameters and buffers.

Warning

Currently state_dict() also accepts positional arguments for destination, prefix and keep_vars in order. However, this is being deprecated and keyword arguments will be enforced in future releases.

Warning

Please avoid the use of argument destination as it is not designed for end-users.

Parameters:

• destination (dict, optional) – If provided, the state of module will be 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 buffer names to compose the keys in state_dict. Default: ''.

• keep_vars (bool, optional) – by default the Tensor s returned in the state dict are detached from autograd. If it’s set to True, detaching will not be performed. Default: False.

Returns:

a dictionary containing a whole state of the module

Return type:

dict

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> module.state_dict().keys()
['bias', 'weight']

to(*args, **kwargs)

Move and/or cast 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 complex dtypes. In addition, this method will only cast the floating point or complex parameters and buffers to dtype (if given). The integral parameters and buffers will be moved device, if that is given, but with dtypes unchanged. When non_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 module

• dtype (torch.dtype) – the desired floating point or complex dtype of the parameters and buffers in this module

• tensor (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:

Module

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: int | str | device | None, recurse: bool = True) → T

Move 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:

Module

train(mode: bool = True) → T

Set the module in training 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.

Parameters:

mode (bool) – whether to set training mode (True) or evaluation mode (False). Default: True.

Returns:

self

Return type:

Module

training: bool

type(dst_type: dtype | str) → T

Casts all parameters and buffers to dst_type.

Note

This method modifies the module in-place.

Parameters:

dst_type (type or string) – the desired type

Returns:

self

Return type:

Module

xpu(device: int | device | None = None) → T

Move 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.

Parameters:

device (int, optional) – if specified, all parameters will be copied to that device

Returns:

self

Return type:

Module

zero_grad(set_to_none: bool = True) → None

Reset 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.

class lzero.model.stochastic_muzero_model.StraightThroughEstimator

Bases: Module

T_destination = ~T_destination

__init__()

Initialize internal Module state, shared by both nn.Module and ScriptModule.

_apply(fn, recurse=True)

_backward_hooks: Dict[int, Callable]

_backward_pre_hooks: Dict[int, Callable]

_buffers: Dict[str, Tensor | None]

_call_impl(*args, **kwargs)

_compiled_call_impl: Callable | None = None

_forward_hooks: Dict[int, Callable]

_forward_hooks_always_called: Dict[int, bool]

_forward_hooks_with_kwargs: Dict[int, bool]

_forward_pre_hooks: Dict[int, Callable]

_forward_pre_hooks_with_kwargs: Dict[int, bool]

_get_backward_hooks()

Return the backward hooks for use in the call function.

It returns two lists, one with the full backward hooks and one with the non-full backward hooks.

_get_backward_pre_hooks()

_get_name()

_is_full_backward_hook: bool | None

_load_from_state_dict(state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs)

Copy parameters and buffers from state_dict into only this module, but not its descendants.

This is called on every submodule in load_state_dict(). Metadata saved for this module in input state_dict is provided as local_metadata. For state dicts without metadata, local_metadata is empty. Subclasses can achieve class-specific backward compatible loading using the version number at local_metadata.get(“version”, None). Additionally, local_metadata can also contain the key assign_to_params_buffers that indicates whether keys should be assigned their corresponding tensor in the state_dict.

Note

state_dict is not the same object as the input state_dict to load_state_dict(). So it can be modified.

Parameters:

• state_dict (dict) – a dict containing parameters and persistent buffers.

• prefix (str) – the prefix for parameters and buffers used in this module

• local_metadata (dict) – a dict containing the metadata for this module. See

• strict (bool) – whether to strictly enforce that the keys in state_dict with prefix match the names of parameters and buffers in this module

• missing_keys (list of str) – if strict=True, add missing keys to this list

• unexpected_keys (list of str) – if strict=True, add unexpected keys to this list

• error_msgs (list of str) – error messages should be added to this list, and will be reported together in load_state_dict()

_load_state_dict_post_hooks: Dict[int, Callable]

_load_state_dict_pre_hooks: Dict[int, Callable]

_maybe_warn_non_full_backward_hook(inputs, result, grad_fn)

_modules: Dict[str, 'Module' | None]

_named_members(get_members_fn, prefix='', recurse=True, remove_duplicate: bool = True)

Help yield various names + members of modules.

_non_persistent_buffers_set: Set[str]

_parameters: Dict[str, Parameter | None]

_register_load_state_dict_pre_hook(hook, with_module=False)

See register_load_state_dict_pre_hook() for details.

A subtle difference is that if with_module is set to False, then the hook will not take the module as the first argument whereas register_load_state_dict_pre_hook() always takes the module as the first argument.

Parameters:

• hook (Callable) – Callable hook that will be invoked before loading the state dict.

• with_module (bool, optional) – Whether or not to pass the module instance to the hook as the first parameter.

_register_state_dict_hook(hook)

Register a post-hook for the state_dict() method.

It should have the following signature::

hook(module, state_dict, prefix, local_metadata) -> None or state_dict

The registered hooks can modify the state_dict inplace or return a new one. If a new state_dict is returned, it will only be respected if it is the root module that state_dict() is called from.

_replicate_for_data_parallel()

_save_to_state_dict(destination, prefix, keep_vars)

Save module state to the destination dictionary.

The destination dictionary will contain the state of the module, but not its descendants. This is called on every submodule in state_dict().

In rare cases, subclasses can achieve class-specific behavior by overriding this method with custom logic.

Parameters:

• destination (dict) – a dict where state will be stored

• prefix (str) – the prefix for parameters and buffers used in this module

_slow_forward(*input, **kwargs)

_state_dict_hooks: Dict[int, Callable]

_state_dict_pre_hooks: Dict[int, Callable]

_version: int = 1

This allows better BC support for load_state_dict(). In state_dict(), the version number will be saved as in the attribute _metadata of the returned state dict, and thus pickled. _metadata is a dictionary 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 shall be bumped, and the module’s _load_from_state_dict method can compare the version number and do appropriate changes if the state dict is from before the change.

_wrapped_call_impl(*args, **kwargs)

add_module(name: str, module: Module | None) → None

Add a child module to the current module.

The module can be accessed as an attribute using the given name.

Parameters:

• name (str) – name of the child module. The child module can be accessed from this module using the given name

• module (Module) – child module to be added to the module.

apply(fn: Callable[[Module], None]) → T

Apply fn recursively to every submodule (as returned by .children()) as well as self.

Typical use includes initializing the parameters of a model (see also nn-init-doc).

Parameters:

fn (Module -> None) – function to be applied to each submodule

Returns:

self

Return type:

Module

Example:

>>> @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_features=2, out_features=2, bias=True)
Parameter containing:
tensor([[1., 1.],
        [1., 1.]], requires_grad=True)
Linear(in_features=2, out_features=2, bias=True)
Parameter containing:
tensor([[1., 1.],
        [1., 1.]], requires_grad=True)
Sequential(
  (0): Linear(in_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
)

bfloat16() → T

Casts all floating point parameters and buffers to bfloat16 datatype.

Note

This method modifies the module in-place.

Returns:

self

Return type:

Module

buffers(recurse: bool = True) → Iterator[Tensor]

Return an iterator over module buffers.

Parameters:

recurse (bool) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are direct members of this module.

Yields:

torch.Tensor – module buffer

Example:

>>> # 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)

call_super_init: bool = False

children() → Iterator[Module]

Return 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

Move all model parameters and buffers to the CPU.

Note

This method modifies the module in-place.

Returns:

self

Return type:

Module

cuda(device: int | device | None = None) → T

Move 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.

Parameters:

device (int, optional) – if specified, all parameters will be copied to that device

Returns:

self

Return type:

Module

double() → T

Casts all floating point parameters and buffers to double datatype.

Note

This method modifies the module in-place.

Returns:

self

Return type:

Module

dump_patches: bool = False

eval() → T

Set 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-disable-grad-doc for a comparison between .eval() and several similar mechanisms that may be confused with it.

Returns:

self

Return type:

Module

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:

Module

forward(x)

Overview:

Forward method for the StraightThroughEstimator. This applies the one-hot argmax function to the input tensor.

Parameters:

x (-) – Input tensor.

Returns:

Transformed tensor after applying one-hot argmax.

Return type:

• (torch.Tensor)

get_buffer(target: str) → Tensor

Return the buffer given by target if it exists, otherwise throw an error.

See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target.

Parameters:

target – The fully-qualified string name of the buffer to look for. (See get_submodule for how to specify a fully-qualified string.)

Returns:

The buffer referenced by target

Return type:

torch.Tensor

Raises:

AttributeError – If the target string references an invalid path or resolves to something that is not a buffer

get_extra_state() → Any

Return 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:

object

get_parameter(target: str) → Parameter

Return the parameter given by target if it exists, otherwise throw an error.

See the docstring for get_submodule for a more detailed explanation of this method’s functionality as well as how to correctly specify target.

Parameters:

target – The fully-qualified string name of the Parameter to look for. (See get_submodule for how to specify a fully-qualified string.)

Returns:

The Parameter referenced by target

Return type:

torch.nn.Parameter

Raises:

AttributeError – If the target string references an invalid path or resolves to something that is not an nn.Parameter

get_submodule(target: str) → Module

Return the submodule given by target if it exists, otherwise throw 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 submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.)

To check whether or not we have the linear submodule, we would call get_submodule("net_b.linear"). To check whether we have the conv submodule, we would call get_submodule("net_b.net_c.conv").

The runtime of get_submodule is bounded by the degree of module nesting in target. A query against named_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:

torch.nn.Module

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:

Module

ipu(device: int | device | None = None) → T

Move 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.

Parameters:

device (int, optional) – if specified, all parameters will be copied to that device

Returns:

self

Return type:

Module

load_state_dict(state_dict: Mapping[str, Any], strict: bool = True, assign: bool = False)

Copy parameters and buffers from state_dict into this module and its descendants.

If strict is True, then the keys of state_dict must exactly match the keys returned by this module’s state_dict() function.

Warning

If assign is True the optimizer must be created after the call to load_state_dict unless get_swap_module_params_on_conversion() is True.

Parameters:

• state_dict (dict) – a dict containing parameters and persistent buffers.

• strict (bool, optional) – whether to strictly enforce that the keys in state_dict match the keys returned by this module’s state_dict() function. Default: True

• assign (bool, optional) – When False, the properties of the tensors in the current module are preserved while when True, the properties of the Tensors in the state dict are preserved. The only exception is the requires_grad field of Default: ``False`

Returns:

• missing_keys is a list of str containing any keys that are expected

  by this module but missing from the provided state_dict.

• unexpected_keys is a list of str containing the keys that are not

  expected by this module but present in the provided state_dict.

Return type:

NamedTuple with missing_keys and unexpected_keys fields

Note

If a parameter or buffer is registered as None and its corresponding key exists in state_dict, load_state_dict() will raise a RuntimeError.

modules() → Iterator[Module]

Return an iterator over all modules in the network.

Yields:

Module – a module in the network

Note

Duplicate modules are returned only once. In the following example, 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_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
)
1 -> Linear(in_features=2, out_features=2, bias=True)

mtia(device: int | device | None = None) → T

Move all model parameters and buffers to the MTIA.

This also makes associated parameters and buffers different objects. So it should be called before constructing optimizer if the module will live on MTIA while being optimized.

Note

This method modifies the module in-place.

Parameters:

device (int, optional) – if specified, all parameters will be copied to that device

Returns:

self

Return type:

Module

named_buffers(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[Tuple[str, Tensor]]

Return an iterator over module buffers, yielding both the name of the buffer as well as the buffer itself.

Parameters:

• prefix (str) – prefix to prepend to all buffer names.

• recurse (bool, optional) – if True, then yields buffers of this module and all submodules. Otherwise, yields only buffers that are 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 buffer

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, buf in self.named_buffers():
>>>     if name in ['running_var']:
>>>         print(buf.size())

named_children() → Iterator[Tuple[str, Module]]

Return an iterator over immediate children modules, yielding both the name of the module as well as the module itself.

Yields:

(str, Module) – Tuple containing a name and child module

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, module in model.named_children():
>>>     if name in ['conv4', 'conv5']:
>>>         print(module)

named_modules(memo: Set[Module] | None = None, prefix: str = '', remove_duplicate: bool = True)

Return an iterator over all modules in the network, yielding both the name of the module as well as the module itself.

Parameters:

• memo – a memo to store the set of modules already added to the result

• prefix – a prefix that will be added to the name of the module

• remove_duplicate – whether to remove the duplicated module instances in the result or not

Yields:

(str, Module) – Tuple of name and module

Note

Duplicate modules are returned only once. In the following example, 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_features=2, out_features=2, bias=True)
  (1): Linear(in_features=2, out_features=2, bias=True)
))
1 -> ('0', Linear(in_features=2, out_features=2, bias=True))

named_parameters(prefix: str = '', recurse: bool = True, remove_duplicate: bool = True) → Iterator[Tuple[str, Parameter]]

Return an iterator over module parameters, yielding both the name of the parameter as well as the parameter itself.

Parameters:

• prefix (str) – prefix to prepend to all parameter names.

• recurse (bool) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module.

• remove_duplicate (bool, optional) – whether to remove the duplicated parameters in the result. Defaults to True.

Yields:

(str, Parameter) – Tuple containing the name and parameter

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> for name, param in self.named_parameters():
>>>     if name in ['bias']:
>>>         print(param.size())

parameters(recurse: bool = True) → Iterator[Parameter]

Return an iterator over module parameters.

This is typically passed to an optimizer.

Parameters:

recurse (bool) – if True, then yields parameters of this module and all submodules. Otherwise, yields only parameters that are direct members of this module.

Yields:

Parameter – module parameter

Example:

>>> # 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)

register_backward_hook(hook: Callable[[Module, Tuple[Tensor, ...] | Tensor, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor]) → RemovableHandle

Register 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_buffer(name: str, tensor: Tensor | None, persistent: bool = True) → None

Add a buffer to the module.

This is typically used to register a buffer that should not to be considered a model parameter. For example, BatchNorm’s running_mean is not a parameter, but is part of the module’s state. Buffers, by default, are persistent and will be saved alongside parameters. This behavior can be changed by setting persistent to False. The only difference between a persistent buffer and a non-persistent buffer is that the latter will not be a part of this module’s state_dict.

Buffers can be accessed as attributes using given names.

Parameters:

• name (str) – name of the buffer. The buffer can be accessed from this module using the given name

• tensor (Tensor or None) – buffer to be registered. If None, then operations that run on buffers, such as cuda, are ignored. If None, the buffer is not included in the module’s state_dict.

• persistent (bool) – whether the buffer is part of this module’s state_dict.

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> self.register_buffer('running_mean', torch.zeros(num_features))

register_forward_hook(hook: Callable[[T, Tuple[Any, ...], Any], Any | None] | Callable[[T, Tuple[Any, ...], Dict[str, Any], Any], Any | None], *, prepend: bool = False, with_kwargs: bool = False, always_call: bool = False) → RemovableHandle

Register a forward hook on the module.

The hook will be called every time after forward() has computed an output.

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 the forward. The hook can modify the output. It can modify the input inplace but it will not have effect on forward since this is called after forward() is called. The hook should have the following signature:

hook(module, args, output) -> None or modified output

If with_kwargs is True, the forward hook will be passed the kwargs 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 provided hook will be fired before all existing forward hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing forward hooks on this torch.nn.modules.Module. Note that global forward hooks registered with register_module_forward_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

• always_call (bool) – If True the hook 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: Callable[[T, Tuple[Any, ...]], Any | None] | Callable[[T, Tuple[Any, ...], Dict[str, Any]], Tuple[Any, Dict[str, Any]] | None], *, prepend: bool = False, with_kwargs: bool = False) → RemovableHandle

Register 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 the forward. 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 existing forward_pre hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing forward_pre hooks on this torch.nn.modules.Module. Note that global forward_pre hooks registered with register_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, Tuple[Tensor, ...] | Tensor, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle

Register 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 and grad_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 of grad_input in subsequent computations. grad_input will only correspond to the inputs given as positional arguments and all kwarg arguments are ignored. Entries in grad_input and grad_output will be None 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 existing backward hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing backward hooks on this torch.nn.modules.Module. Note that global backward hooks registered with register_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, Tuple[Tensor, ...] | Tensor], None | Tuple[Tensor, ...] | Tensor], prepend: bool = False) → RemovableHandle

Register 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 of grad_output in subsequent computations. Entries in grad_output will be None 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 existing backward_pre hooks on this torch.nn.modules.Module. Otherwise, the provided hook will be fired after all existing backward_pre hooks on this torch.nn.modules.Module. Note that global backward_pre hooks registered with register_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)

Register 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 the incompatible_keys argument is a NamedTuple consisting of attributes missing_keys and unexpected_keys. missing_keys is a list of str containing the missing keys and unexpected_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 load_state_dict() with strict=True are affected by modifications the hook makes to missing_keys or unexpected_keys, as expected. Additions to either set of keys will result in an error being thrown when strict=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_load_state_dict_pre_hook(hook)

Register a pre-hook to be run before module’s load_state_dict() is called.

It should have the following signature::

hook(module, state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs) -> None # noqa: B950

Parameters:

hook (Callable) – Callable hook that will be invoked before loading the state dict.

register_module(name: str, module: Module | None) → None

Alias for add_module().

register_parameter(name: str, param: Parameter | None) → None

Add a parameter to the module.

The parameter can be accessed as an attribute using given name.

Parameters:

• name (str) – name of the parameter. The parameter can be accessed from this module using the given name

• param (Parameter or None) – parameter to be added to the module. If None, then operations that run on parameters, such as cuda, are ignored. If None, the parameter is not included in the module’s state_dict.

register_state_dict_post_hook(hook)

Register a post-hook for the state_dict() method.

It should have the following signature::

hook(module, state_dict, prefix, local_metadata) -> None

The registered hooks can modify the state_dict inplace.

register_state_dict_pre_hook(hook)

Register a pre-hook for the state_dict() method.

It should have the following signature::

hook(module, prefix, keep_vars) -> None

The registered hooks can be used to perform pre-processing before the state_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-disable-grad-doc for a comparison between .requires_grad_() and several similar mechanisms that may be confused with it.

Parameters:

requires_grad (bool) – whether autograd should record operations on parameters in this module. Default: True.

Returns:

self

Return type:

Module

set_extra_state(state: Any) → None

Set extra state contained in the loaded state_dict.

This function is called from load_state_dict() to handle any extra state found within the state_dict. Implement this function and a corresponding get_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

set_submodule(target: str, module: Module) → None

Set the submodule given by target if it exists, otherwise throw 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 submodule net_b, which itself has two submodules net_c and linear. net_c then has a submodule conv.)

To overide the Conv2d with a new submodule Linear, you would call set_submodule("net_b.net_c.conv", nn.Linear(33, 16)).

Parameters:

• target – The fully-qualified string name of the submodule to look for. (See above example for how to specify a fully-qualified string.)

• module – The module to set the submodule to.

Raises:

• ValueError – If the target string is empty

• AttributeError – If the target string references an invalid path or resolves to something that is not an nn.Module

share_memory() → T

See torch.Tensor.share_memory_().

state_dict(*args, destination=None, prefix='', keep_vars=False)

Return a dictionary containing references to the whole state of the module.

Both parameters and persistent buffers (e.g. running averages) are included. 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 references to the module’s parameters and buffers.

Warning

Currently state_dict() also accepts positional arguments for destination, prefix and keep_vars in order. However, this is being deprecated and keyword arguments will be enforced in future releases.

Warning

Please avoid the use of argument destination as it is not designed for end-users.

Parameters:

• destination (dict, optional) – If provided, the state of module will be 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 buffer names to compose the keys in state_dict. Default: ''.

• keep_vars (bool, optional) – by default the Tensor s returned in the state dict are detached from autograd. If it’s set to True, detaching will not be performed. Default: False.

Returns:

a dictionary containing a whole state of the module

Return type:

dict

Example:

>>> # xdoctest: +SKIP("undefined vars")
>>> module.state_dict().keys()
['bias', 'weight']

to(*args, **kwargs)

Move and/or cast 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 complex dtypes. In addition, this method will only cast the floating point or complex parameters and buffers to dtype (if given). The integral parameters and buffers will be moved device, if that is given, but with dtypes unchanged. When non_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 module

• dtype (torch.dtype) – the desired floating point or complex dtype of the parameters and buffers in this module

• tensor (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:

Module

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: int | str | device | None, recurse: bool = True) → T

Move 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:

Module

train(mode: bool = True) → T

Set the module in training 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.

Parameters:

mode (bool) – whether to set training mode (True) or evaluation mode (False). Default: True.

Returns:

self

Return type:

Module

training: bool

type(dst_type: dtype | str) → T

Casts all parameters and buffers to dst_type.

Note

This method modifies the module in-place.

Parameters:

dst_type (type or string) – the desired type

Returns:

self

Return type:

Module

xpu(device: int | device | None = None) → T

Move 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.

Parameters:

device (int, optional) – if specified, all parameters will be copied to that device

Returns:

self

Return type:

Module

zero_grad(set_to_none: bool = True) → None

Reset 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.

class lzero.model.stochastic_muzero_model.OnehotArgmax(*args, **kwargs)

Bases: Function

Overview:

Custom PyTorch function for one-hot argmax. This function transforms the input tensor into a one-hot tensor where the index with the maximum value in the original tensor is set to 1 and all other indices are set to 0. It allows gradients to flow to the encoder during backpropagation.

For more information, refer to: https://pytorch.org/tutorials/beginner/examples_autograd/two_layer_net_custom_function.html

__init__(*args, **kwargs)

_backward_cls

alias of OnehotArgmaxBackward

_compiled_autograd_backward_state

static _compiled_autograd_key(ctx)

_get_compiled_autograd_symints()

_input_metadata

_is_compiled_autograd_tracing()

_materialize_non_diff_grads

_raw_saved_tensors

static _register_hook(backward_hooks, hook)

_register_hook_dict()

_sequence_nr()

_set_sequence_nr()

classmethod apply(*args, **kwargs)

static backward(ctx, grad_output)

Overview:

Backward method for the one-hot argmax function. This method allows gradients to flow to the encoder during backpropagation.

Parameters:

• ctx (-) – A context object that was stashed in the forward pass.

• grad_output (-) – The gradient of the output tensor.

Returns:

The gradient of the input tensor.

Return type:

• (torch.Tensor)

dirty_tensors

static forward(ctx, input)

Overview:

Forward method for the one-hot argmax function. This method transforms the input tensor into a one-hot tensor.

Parameters:

• ctx (-) – A context object that can be used to stash information for

• computation. (backward) –

• input (-) – Input tensor.

Returns:

One-hot tensor.

Return type:

• (torch.Tensor)

generate_vmap_rule = False

static jvp(ctx: Any, *grad_inputs: Any) → Any

Define a formula for differentiating the operation with forward mode automatic differentiation.

This function is to be overridden by all subclasses. It must accept a context ctx as the first argument, followed by as many inputs as the forward() got (None will be passed in for non tensor inputs of the forward function), and it should return as many tensors as there were outputs to forward(). Each argument is the gradient w.r.t the given input, and each returned value should be the gradient w.r.t. the corresponding output. If an output is not a Tensor or the function is not differentiable with respect to that output, you can just pass None as a gradient for that input.

You can use the ctx object to pass any value from the forward to this functions.

mark_dirty(*args: Tensor)

Mark given tensors as modified in an in-place operation.

This should be called at most once, in either the setup_context() or forward() methods, and all arguments should be inputs.

Every tensor that’s been modified in-place in a call to forward() should be given to this function, to ensure correctness of our checks. It doesn’t matter whether the function is called before or after modification.

Examples::

>>> # xdoctest: +REQUIRES(env:TORCH_DOCTEST_AUTOGRAD)
>>> class Inplace(Function):
>>>     @staticmethod
>>>     def forward(ctx, x):
>>>         x_npy = x.numpy() # x_npy shares storage with x
>>>         x_npy += 1
>>>         ctx.mark_dirty(x)
>>>         return x
>>>
>>>     @staticmethod
>>>     @once_differentiable
>>>     def backward(ctx, grad_output):
>>>         return grad_output
>>>
>>> a = torch.tensor(1., requires_grad=True, dtype=torch.double).clone()
>>> b = a * a
>>> Inplace.apply(a)  # This would lead to wrong gradients!
>>>                   # but the engine would not know unless we mark_dirty
>>> # xdoctest: +SKIP
>>> b.backward() # RuntimeError: one of the variables needed for gradient
>>>              # computation has been modified by an inplace operation

mark_non_differentiable(*args: Tensor)

Mark outputs as non-differentiable.

This should be called at most once, in either the setup_context() or forward() methods, and all arguments should be tensor outputs.

This will mark outputs as not requiring gradients, increasing the efficiency of backward computation. You still need to accept a gradient for each output in backward(), but it’s always going to be a zero tensor with the same shape as the shape of a corresponding output.

This is used e.g. for indices returned from a sort. See example::

>>> class Func(Function):
>>>     @staticmethod
>>>     def forward(ctx, x):
>>>         sorted, idx = x.sort()
>>>         ctx.mark_non_differentiable(idx)
>>>         ctx.save_for_backward(x, idx)
>>>         return sorted, idx
>>>
>>>     @staticmethod
>>>     @once_differentiable
>>>     def backward(ctx, g1, g2):  # still need to accept g2
>>>         x, idx = ctx.saved_tensors
>>>         grad_input = torch.zeros_like(x)
>>>         grad_input.index_add_(0, idx, g1)
>>>         return grad_input

mark_shared_storage(*pairs)

materialize_grads

maybe_clear_saved_tensors()

metadata

name()

needs_input_grad

next_functions

non_differentiable

register_hook()

register_prehook()

requires_grad

save_for_backward(*tensors: Tensor)

Save given tensors for a future call to backward().

save_for_backward should be called at most once, in either the setup_context() or forward() methods, and only with tensors.

All tensors intended to be used in the backward pass should be saved with save_for_backward (as opposed to directly on ctx) to prevent incorrect gradients and memory leaks, and enable the application of saved tensor hooks. See torch.autograd.graph.saved_tensors_hooks.

Note that if intermediary tensors, tensors that are neither inputs nor outputs of forward(), are saved for backward, your custom Function may not support double backward. Custom Functions that do not support double backward should decorate their backward() method with @once_differentiable so that performing double backward raises an error. If you’d like to support double backward, you can either recompute intermediaries based on the inputs during backward or return the intermediaries as the outputs of the custom Function. See the double backward tutorial for more details.

In backward(), saved tensors can be accessed through the saved_tensors attribute. Before returning them to the user, a check is made to ensure they weren’t used in any in-place operation that modified their content.

Arguments can also be None. This is a no-op.

See extending-autograd for more details on how to use this method.

Example::

>>> # xdoctest: +REQUIRES(env:TORCH_DOCTEST_AUTOGRAD)
>>> class Func(Function):
>>>     @staticmethod
>>>     def forward(ctx, x: torch.Tensor, y: torch.Tensor, z: int):
>>>         w = x * z
>>>         out = x * y + y * z + w * y
>>>         ctx.save_for_backward(x, y, w, out)
>>>         ctx.z = z  # z is not a tensor
>>>         return out
>>>
>>>     @staticmethod
>>>     @once_differentiable
>>>     def backward(ctx, grad_out):
>>>         x, y, w, out = ctx.saved_tensors
>>>         z = ctx.z
>>>         gx = grad_out * (y + y * z)
>>>         gy = grad_out * (x + z + w)
>>>         gz = None
>>>         return gx, gy, gz
>>>
>>> a = torch.tensor(1., requires_grad=True, dtype=torch.double)
>>> b = torch.tensor(2., requires_grad=True, dtype=torch.double)
>>> c = 4
>>> d = Func.apply(a, b, c)

save_for_forward(*tensors: Tensor)

Save given tensors for a future call to jvp().

save_for_forward should be called at most once, in either the setup_context() or forward() methods, and all arguments should be tensors.

In jvp(), saved objects can be accessed through the saved_tensors attribute.

Arguments can also be None. This is a no-op.

See extending-autograd for more details on how to use this method.

Example::

>>> # xdoctest: +SKIP
>>> class Func(torch.autograd.Function):
>>>     @staticmethod
>>>     def forward(ctx, x: torch.Tensor, y: torch.Tensor, z: int):
>>>         ctx.save_for_backward(x, y)
>>>         ctx.save_for_forward(x, y)
>>>         ctx.z = z
>>>         return x * y * z
>>>
>>>     @staticmethod
>>>     def jvp(ctx, x_t, y_t, _):
>>>         x, y = ctx.saved_tensors
>>>         z = ctx.z
>>>         return z * (y * x_t + x * y_t)
>>>
>>>     @staticmethod
>>>     def vjp(ctx, grad_out):
>>>         x, y = ctx.saved_tensors
>>>         z = ctx.z
>>>         return z * grad_out * y, z * grad_out * x, None
>>>
>>>     a = torch.tensor(1., requires_grad=True, dtype=torch.double)
>>>     t = torch.tensor(1., dtype=torch.double)
>>>     b = torch.tensor(2., requires_grad=True, dtype=torch.double)
>>>     c = 4
>>>
>>>     with fwAD.dual_level():
>>>         a_dual = fwAD.make_dual(a, t)
>>>         d = Func.apply(a_dual, b, c)

saved_for_forward

saved_tensors

saved_variables

set_materialize_grads(value: bool)

Set whether to materialize grad tensors. Default is True.

This should be called only from either the setup_context() or forward() methods.

If True, undefined grad tensors will be expanded to tensors full of zeros prior to calling the backward() and jvp() methods.

Example::

>>> # xdoctest: +REQUIRES(env:TORCH_DOCTEST_AUTOGRAD)
>>> class SimpleFunc(Function):
>>>     @staticmethod
>>>     def forward(ctx, x):
>>>         return x.clone(), x.clone()
>>>
>>>     @staticmethod
>>>     @once_differentiable
>>>     def backward(ctx, g1, g2):
>>>         return g1 + g2  # No check for None necessary
>>>
>>> # We modify SimpleFunc to handle non-materialized grad outputs
>>> class Func(Function):
>>>     @staticmethod
>>>     def forward(ctx, x):
>>>         ctx.set_materialize_grads(False)
>>>         ctx.save_for_backward(x)
>>>         return x.clone(), x.clone()
>>>
>>>     @staticmethod
>>>     @once_differentiable
>>>     def backward(ctx, g1, g2):
>>>         x, = ctx.saved_tensors
>>>         grad_input = torch.zeros_like(x)
>>>         if g1 is not None:  # We must check for None now
>>>             grad_input += g1
>>>         if g2 is not None:
>>>             grad_input += g2
>>>         return grad_input
>>>
>>> a = torch.tensor(1., requires_grad=True)
>>> b, _ = Func.apply(a)  # induces g2 to be undefined

static setup_context(ctx: Any, inputs: Tuple[Any, ...], output: Any) → Any

There are two ways to define the forward pass of an autograd.Function.

Either:

1. Override forward with the signature forward(ctx, *args, **kwargs). setup_context is not overridden. Setting up the ctx for backward happens inside the forward.

2. Override forward with the signature forward(*args, **kwargs) and override setup_context. Setting up the ctx for backward happens inside setup_context (as opposed to inside the forward)

See torch.autograd.Function.forward() and extending-autograd for more details.

to_save

static vjp(ctx: Any, *grad_outputs: Any) → Any

Define a formula for differentiating the operation with backward mode automatic differentiation.

This function is to be overridden by all subclasses. (Defining this function is equivalent to defining the vjp function.)

It must accept a context ctx as the first argument, followed by as many outputs as the forward() returned (None will be passed in for non tensor outputs of the forward function), and it should return as many tensors, as there were inputs to forward(). Each argument is the gradient w.r.t the given output, and each returned value should be the gradient w.r.t. the corresponding input. If an input is not a Tensor or is a Tensor not requiring grads, you can just pass None as a gradient for that input.

The context can be used to retrieve tensors saved during the forward pass. It also has an attribute ctx.needs_input_grad as a tuple of booleans representing whether each input needs gradient. E.g., backward() will have ctx.needs_input_grad[0] = True if the first input to forward() needs gradient computed w.r.t. the output.

static vmap(info, in_dims, *args)

Define the behavior for this autograd.Function underneath torch.vmap().

For a torch.autograd.Function() to support torch.vmap(), you must either override this static method, or set generate_vmap_rule to True (you may not do both).

If you choose to override this staticmethod: it must accept

• an info object as the first argument. info.batch_size specifies the size of the dimension being vmapped over, while info.randomness is the randomness option passed to torch.vmap().

• an in_dims tuple as the second argument. For each arg in args, in_dims has a corresponding Optional[int]. It is None if the arg is not a Tensor or if the arg is not being vmapped over, otherwise, it is an integer specifying what dimension of the Tensor is being vmapped over.

• *args, which is the same as the args to forward().

The return of the vmap staticmethod is a tuple of (output, out_dims). Similar to in_dims, out_dims should be of the same structure as output and contain one out_dim per output that specifies if the output has the vmapped dimension and what index it is in.

Please see func-autograd-function for more details.