netket.nn.FastMaskedConv1D
netket.nn.FastMaskedConv1D#
- class netket.nn.FastMaskedConv1D[source]#
Bases:
flax.linen.module.Module
1D convolution module with mask for fast autoregressive NN.
The fast autoregressive sampling is described in Ramachandran et. {it al}. To generate one sample using an autoregressive network, we need to evaluate the network N times, where N is the number of input sites. But we only change one input site each time, so we can cache unchanged intermediate results and avoid repeated computation.
- Inheritance
- Attributes
- feature_group_count: int = 1#
1).
- Type
if specified, divides the input features into groups (default
- parent: Optional[Union[Type[flax.linen.module.Module], Type[flax.core.scope.Scope], Type[flax.linen.module._Sentinel]]] = <flax.linen.module._Sentinel object>#
- precision: Any = None#
numerical precision of the computation, see jax.lax.Precision for details.
- scope = None#
- Methods
- __call__(inputs)[source]#
Applies the masked convolution to all input sites.
- Parameters
inputs (
Union
[ndarray
,DeviceArray
,Tracer
]) β input data with dimensions (batch, size, features).- Return type
Union
[ndarray
,DeviceArray
,Tracer
]- Returns
The convolved data.
- apply(variables, *args, rngs=None, method=None, mutable=False, capture_intermediates=False, **kwargs)#
Applies a module method to variables and returns output and modified variables.
Note that method should be set if one would like to call apply on a different class method than
__call__
. For instance, suppose a Transformer modules has a method called encode, then the following calls apply on that method:model = Transformer() encoded = model.apply({'params': params}, x, method=Transformer.encode)
If a function instance is provided, the unbound function is used. For instance, the example below is equivalent to the one above:
encoded = model.apply({'params': params}, x, method=model.encode)
Note
method
can also be a function that is not defined inTransformer
. In that case, the function should have at least one argument representing an instance of the Module class:def other_fn(instance, ...): instance.some_module_attr(...) ... model.apply({'params': params}, x, method=other_fn)
- Parameters
variables (
Mapping
[str
,Mapping
[str
,Any
]]) β A dictionary containing variables keyed by variable collections. Seeflax.core.variables
for more details about variables.*args β Named arguments passed to the specified apply method.
rngs (
Optional
[Dict
[str
,Any
]]) β a dict of PRNGKeys to initialize the PRNG sequences. The βparamsβ PRNG sequence is used to initialize parameters.method (
Optional
[Callable
[...
,Any
]]) β A function to call apply on. This is generally a function in the module. If provided, applies this method. If not provided, applies the__call__
method of the module.mutable (
Union
[bool
,str
,Collection
[str
],DenyList
]) β Can be bool, str, or list. Specifies which collections should be treated as mutable:bool
: all/no collections are mutable.str
: The name of a single mutable collection.list
: A list of names of mutable collections.capture_intermediates (
Union
[bool
,Callable
[[Module
,str
],bool
]]) β If True, captures intermediate return values of all Modules inside the βintermediatesβ collection. By default only the return values of all__call__
methods are stored. A function can be passed to change the filter behavior. The filter function takes the Module instance and method name and returns a bool indicating whether the output of that method invocation should be stored.**kwargs β Keyword arguments passed to the specified apply method.
- Return type
- Returns
If
mutable
is False, returns output. If any collections are mutable, returns(output, vars)
, wherevars
are is a dict of the modified collections.
- bias_init(shape, dtype=<class 'jax.numpy.float64'>)#
An initializer that returns a constant array full of zeros.
The
key
argument is ignored.>>> import jax, jax.numpy as jnp >>> jax.nn.initializers.zeros(jax.random.PRNGKey(42), (2, 3), jnp.float32) DeviceArray([[0., 0., 0.], [0., 0., 0.]], dtype=float32)
- Parameters
dtype (Any) β
- bind(variables, *args, rngs=None, mutable=False)#
Creates an interactive Module instance by binding variables and RNGs.
bind
provides an βinteractiveβ instance of a Module directly without transforming a function withapply
. This is particalary useful for debugging and interactive use cases like notebooks where a function would limit the ability split up code into different cells.Once the variables (and optionally RNGs) are bound to a
Module
it becomes a stateful object. Note that idiomatic JAX is functional and therefore an interactive instance does not mix well well with vanilla JAX APIs.bind()
should only be used for interactive experimentation, and in all other cases we strongly encourage to useapply()
instead.Example:
import jax import jax.numpy as jnp import flax.linen as nn class AutoEncoder(nn.Module): def setup(self): self.encoder = nn.Dense(3) self.decoder = nn.Dense(5) def __call__(self, x): return self.decoder(self.encoder(x)) x = jnp.ones((16, 9)) ae = AutoEncoder() variables = ae.init(jax.random.PRNGKey(0), x) model = ae.bind(variables) z = model.encoder(x) x_reconstructed = model.decoder(z)
- Parameters
variables (
Mapping
[str
,Mapping
[str
,Any
]]) β A dictionary containing variables keyed by variable collections. Seeflax.core.variables
for more details about variables.*args β Named arguments (not used).
rngs (
Optional
[Dict
[str
,Any
]]) β a dict of PRNGKeys to initialize the PRNG sequences.mutable (
Union
[bool
,str
,Collection
[str
],DenyList
]) βCan be bool, str, or list. Specifies which collections should be treated as mutable:
bool
: all/no collections are mutable.str
: The name of a single mutable collection.list
: A list of names of mutable collections.
- Returns
A copy of this instance with bound variables and RNGs.
- clone(*, parent=None, **updates)#
Creates a clone of this Module, with optionally updated arguments.
- get_variable(col, name, default=None)#
Retrieves the value of a Variable.
- Parameters
- Return type
TypeVar
(T
)- Returns
The value of the input variable, of the default value if the variable doesnβt exist in this scope.
- has_rng(name)#
Returns true if a PRNGSequence with name name exists.
- has_variable(col, name)#
Checks if a variable of given collection and name exists in this Module.
See
flax.core.variables
for more explanation on variables and collections.
- init(rngs, *args, method=None, mutable=DenyList(deny='intermediates'), **kwargs)#
Initializes a module method with variables and returns modified variables.
Jitting init initializes a model lazily using only the shapes of the provided arguments, and avoids computing the forward pass with actual values. Example:
jit_init = jax.jit(SomeModule(...).init) jit_init(rng, jnp.ones(input_shape, jnp.float32))
- Parameters
rngs (
Union
[Any
,Dict
[str
,Any
]]) β The rngs for the variable collections.*args β Named arguments passed to the init function.
method (
Optional
[Callable
[...
,Any
]]) β An optional method. If provided, applies this method. If not provided, applies the__call__
method.mutable (
Union
[bool
,str
,Collection
[str
],DenyList
]) β Can be bool, str, or list. Specifies which collections should be treated as mutable:bool
: all/no collections are mutable.str
: The name of a single mutable collection.list
: A list of names of mutable collections. By default all collections except βintermediatesβ are mutable.**kwargs β Keyword arguments passed to the init function.
- Return type
FrozenDict
[str
,Mapping
[str
,Any
]]- Returns
The initialized variable dict.
- init_with_output(rngs, *args, method=None, mutable=DenyList(deny='intermediates'), **kwargs)#
Initializes a module method with variables and returns output and modified variables.
- Parameters
rngs (
Union
[Any
,Dict
[str
,Any
]]) β The rngs for the variable collections.*args β Named arguments passed to the init function.
method (
Optional
[Callable
[...
,Any
]]) β An optional method. If provided, applies this method. If not provided, applies the__call__
method.mutable (
Union
[bool
,str
,Collection
[str
],DenyList
]) β Can be bool, str, or list. Specifies which collections should be treated as mutable:bool
: all/no collections are mutable.str
: The name of a single mutable collection.list
: A list of names of mutable collections. By default all collections except βintermediatesβ are mutable.**kwargs β Keyword arguments passed to the init function.
- Return type
- Returns
(output, vars)`, where
vars
are is a dict of the modified collections.
- is_mutable_collection(col)#
Returns true if the collection col is mutable.
- kernel_init(shape, dtype=<class 'jax.numpy.float64'>)#
- make_rng(name)#
Returns a new RNG key from a given RNG sequence for this Module.
The new RNG key is split from the previous one. Thus, every call to make_rng returns a new RNG key, while still guaranteeing full reproducibility.
TODO: Link to Flax RNG design note.
- param(name, init_fn, *init_args)#
Declares and returns a parameter in this Module.
Parameters are read-only variables in the collection named βparamsβ. See
flax.core.variables
for more details on variables.The first argument of init_fn is assumed to be a PRNG key, which is provided automatically and does not have to be passed using init_args:
mean = self.param('mean', lecun_normal(), (2, 2))
In the example above, the function lecun_normal expects two arguments: key and shape, but only shape has to be provided explicitly; key is set automatically using the PRNG for params that is passed when initializing the module using
init()
.- Parameters
- Return type
TypeVar
(T
)- Returns
The value of the initialized parameter.
- put_variable(col, name, value)#
Sets the value of a Variable.
- Parameters
Returns:
- setup()#
Initializes a Module lazily (similar to a lazy
__init__
).setup
is called once lazily on a module instance when a module is bound, immediately before any other methods like__call__
are invoked, or before asetup
-defined attribute on self is accessed.This can happen in three cases:
Immediately when invoking
apply()
,init()
orinit_and_output()
.Once the module is given a name by being assigned to an attribute of another module inside the other moduleβs
setup
method (see__setattr__()
):class MyModule(nn.Module): def setup(self): submodule = Conv(...) # Accessing `submodule` attributes does not yet work here. # The following line invokes `self.__setattr__`, which gives # `submodule` the name "conv1". self.conv1 = submodule # Accessing `submodule` attributes or methods is now safe and # either causes setup() to be called once.
Once a module is constructed inside a method wrapped with
compact()
, immediately before another method is called orsetup
defined attribute is accessed.
- Return type
- sow(col, name, value, reduce_fn=<function <lambda>>, init_fn=<function <lambda>>)#
Stores a value in a collection.
Collections can be used to collect intermediate values without the overhead of explicitly passing a container through each Module call.
If the target collection is not mutable sow behaves like a no-op and returns False.
Example:
import jax import jax.numpy as jnp import flax.linen as nn class Foo(nn.Module): @nn.compact def __call__(self, x): h = nn.Dense(4)(x) self.sow('intermediates', 'h', h) return nn.Dense(2)(h) x = jnp.ones((16, 9)) model = Foo() variables = model.init(jax.random.PRNGKey(0), x) y, state = model.apply(variables, x, mutable=['intermediates']) print(state['intermediates']) # {'h': (...,)}
By default the values are stored in a tuple and each stored value is appended at the end. This way all intermediates can be tracked when the same module is called multiple times. Alternatively, a custom init/reduce function can be passed:
class Foo2(nn.Module): @nn.compact def __call__(self, x): init_fn = lambda: 0 reduce_fn = lambda a, b: a + b self.sow('intermediates', 'h', x, init_fn=init_fn, reduce_fn=reduce_fn) self.sow('intermediates', 'h', x * 2, init_fn=init_fn, reduce_fn=reduce_fn) return x model = Foo2() variables = model.init(jax.random.PRNGKey(0), x) y, state = model.apply(variables, jnp.ones((1, 1)), mutable=['intermediates']) print(state['intermediates']) # ==> {'h': [[3.]]}
- Parameters
col (
str
) β The name of the variable collection.name (
str
) β The name of the variable.value (
TypeVar
(T
)) β The value of the variable.reduce_fn (
Callable
[[TypeVar
(K
),TypeVar
(T
)],TypeVar
(K
)]) β The function used to combine the existing value with the new value. The default is to append the value to a tuple.init_fn (
Callable
[[],TypeVar
(K
)]) β For the first value stored, reduce_fn will be passed the result of init_fn together with the value to be stored. The default is an empty tuple.
- Return type
- Returns
True if the value has been stored successfully, False otherwise.
- tabulate(rngs, *args, method=None, mutable=True, depth=None, exclude_methods=(), **kwargs)#
Creates a summary of the Module represented as a table.
This method has the same signature as init, but instead of returning the variables, it returns the string summarizing the Module in a table. tabulate uses jax.eval_shape to run the forward computation without consuming any FLOPs or allocating memory.
Example:
import jax import jax.numpy as jnp import flax.linen as nn class Foo(nn.Module): @nn.compact def __call__(self, x): h = nn.Dense(4)(x) return nn.Dense(2)(h) x = jnp.ones((16, 9)) print(Foo().tabulate(jax.random.PRNGKey(0), x))
This gives the following output:
Foo Summary βββββββββββ³ββββββββββββββββ³βββββββββββββββββββββββ β path β outputs β params β β‘βββββββββββββββββββββββββββββββββββββββββββββββββ© β Inputs β float32[16,9] β β βββββββββββΌββββββββββββββββΌβββββββββββββββββββββββ€ β Dense_0 β float32[16,4] β bias: float32[4] β β β β kernel: float32[9,4] β β β β β β β β 40 (160 B) β βββββββββββΌββββββββββββββββΌβββββββββββββββββββββββ€ β Dense_1 β float32[16,2] β bias: float32[2] β β β β kernel: float32[4,2] β β β β β β β β 10 (40 B) β βββββββββββΌββββββββββββββββΌβββββββββββββββββββββββ€ β Foo β float32[16,2] β β βββββββββββΌββββββββββββββββΌβββββββββββββββββββββββ€ β β Total β 50 (200 B) β βββββββββββ΄ββββββββββββββββ΄βββββββββββββββββββββββ Total Parameters: 50 (200 B)
Note: rows order in the table does not represent execution order, instead it aligns with the order of keys in variables which are sorted alphabetically.
- Parameters
rngs (
Union
[Any
,Dict
[str
,Any
]]) β The rngs for the variable collections.*args β The arguments to the forward computation.
method (
Optional
[Callable
[...
,Any
]]) β An optional method. If provided, applies this method. If not provided, applies the__call__
method.mutable (
Union
[bool
,str
,Collection
[str
],DenyList
]) β Can be bool, str, or list. Specifies which collections should be treated as mutable:bool
: all/no collections are mutable.str
: The name of a single mutable collection.list
: A list of names of mutable collections. By default all collections except βintermediatesβ are mutable.depth (
Optional
[int
]) β controls how many submodule deep the summary can go. By default its None which means no limit. If a submodule is not shown because of the depth limit, its parameter count and bytes will be added to the row of its first shown ancestor such that the sum of all rows always adds up to the total number of parameters of the Module.exclude_methods (
Sequence
[str
]) β A sequence of strings that specifies which methods should be ignored. In case a module calls a helper method from its main method, use this argument to exclude the helper method from the summary to avoid ambiguity.**kwargs β keyword arguments to pass to the forward computation.
- Return type
- Returns
A string summarizing the Module.
- update_site(inputs, index)[source]#
Adds an input site into the cache, and applies the masked convolution to the cache.
- Parameters
inputs (
Union
[ndarray
,DeviceArray
,Tracer
]) β an input site to be added into the cache with dimensions (batch, features).index (
int
) β the index of the output site. The index of the input site should be index - self.exclusive.
- Return type
Union
[ndarray
,DeviceArray
,Tracer
]- Returns
The next output site with dimensions (batch, features).
- variable(col, name, init_fn=None, *init_args)#
Declares and returns a variable in this Module.
See
flax.core.variables
for more information. See alsoparam()
for a shorthand way to define read-only variables in the βparamsβ collection.Contrary to
param()
, all arguments passing using init_fn should be passed on explicitly:key = self.make_rng('stats') mean = self.variable('stats', 'mean', lecun_normal(), key, (2, 2))
In the example above, the function lecun_normal expects two arguments: key and shape, and both have to be passed on. The PRNG for stats has to be provided explicitly when calling
init()
andapply()
.- Parameters
col (
str
) β The variable collection name.name (
str
) β The variable name.init_fn (
Optional
[Callable
[...
,Any
]]) β The function that will be called to compute the initial value of this variable. This function will only be called the first time this variable is used in this module. If None, the variable must already be initialized otherwise an error is raised.*init_args β The arguments to pass to init_fn.
- Return type
- Returns
A
flax.core.variables.Variable
that can be read or set via β.valueβ attribute. Throws an error if the variable exists already.