# Copyright 2021 The NetKet Authors - All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import abc
from typing import Optional, Union, Callable
from collections.abc import Iterator
import numpy as np
from flax import linen as nn
from jax import numpy as jnp
from netket import jax as nkjax
from netket.jax import sharding
from netket import config
from netket.hilbert import AbstractHilbert
from netket.utils import get_afun_if_module, numbers, struct, wrap_afun
from netket.utils.types import PyTree, DType, SeedT
from netket.jax import HashablePartial
[docs]
class SamplerState(struct.Pytree):
"""
Base class holding the state of a sampler.
"""
[docs]
class Sampler(struct.Pytree):
"""
Abstract base class for all samplers.
It contains the fields that all of them should possess, defining the common
API.
Note that fields marked with `pytree_node=False` are treated as static arguments
when jitting.
Subclasses should be NetKet dataclasses and they should define the `_init_state`,
`_reset` and `_sample_chain` methods which only accept positional arguments.
See the respective method's definition for its signature.
Notice that those methods are different from the API-entry point without the leading
underscore in order to allow us to share some pre-processing code between samplers
and simplify the definition of a new sampler.
"""
hilbert: AbstractHilbert = struct.field(pytree_node=False)
"""The Hilbert space to sample."""
machine_pow: int = struct.field(default=2)
"""The power to which the machine should be exponentiated to generate the pdf."""
dtype: DType = struct.field(pytree_node=False, default=float)
"""The dtype of the states sampled."""
[docs]
def __init__(
self,
hilbert: AbstractHilbert,
*,
machine_pow: int = 2,
dtype: DType = float,
):
"""
Construct a Monte Carlo sampler.
Args:
hilbert: The Hilbert space to sample.
machine_pow: The power to which the machine should be exponentiated to generate the pdf (default = 2).
dtype: The dtype of the states sampled (default = np.float64).
"""
# Raise errors if hilbert is not an Hilbert
if not isinstance(hilbert, AbstractHilbert):
raise TypeError(
"\n\nThe argument `hilbert` of a Sampler must be a subtype "
"of netket.hilbert.AbstractHilbert, but you passed in an object "
f"of type {type(hilbert)}, which is not an AbstractHilbert.\n\n"
"TO FIX THIS ERROR,\ndouble check the arguments passed to the "
"sampler when constructing it, and verify that they have the "
"correct types.\n\n"
"For more information, check the correct arguments in the API "
"reference at https://netket.readthedocs.io/en/latest/api/sampler.html"
"\n"
)
if not np.issubdtype(numbers.dtype(machine_pow), np.integer):
raise ValueError(
f"machine_pow ({self.machine_pow}) must be a positive integer"
)
self.hilbert = hilbert
self.machine_pow = machine_pow
self.dtype = dtype
@property
def n_chains_per_rank(self) -> int:
"""
The total number of independent chains per MPI rank (or jax device
if you set `NETKET_EXPERIMENTAL_SHARDING=1`).
If you are not distributing the calculation among different MPI ranks
or jax devices, this is equal to :attr:`~Sampler.n_chains`.
In general this is equal to
.. code:: python
from netket.jax import sharding
sampler.n_chains // sharding.device_count()
"""
n_devices = sharding.device_count()
res, remainder = divmod(self.n_chains, n_devices)
if remainder != 0:
raise RuntimeError(
"The number of chains is not a multiple of the number of mpi ranks"
)
return res
@property
def n_chains(self) -> int:
"""
The total number of independent chains.
This is at least equal to the total number of MPI ranks/jax devices that
are used to distribute the calculation.
"""
# This is the default number of chains, intended for generic non-mcmc
# samplers which don't have a concept of chains.
# We assume there is 1 dummy chain per mpi rank / jax device.
# Currently this is used by the exact samplers (ExactSampler, ARDirectSampler).
return sharding.device_count()
@property
def n_batches(self) -> int:
r"""
The batch size of the configuration $\sigma$ used by this sampler on this
jax process.
This is used to determine the shape of the batches generated in a single process.
This is needed because when using MPI, every process must create a batch of chains
of :attr:`~Sampler.n_chains_per_rank`, while when using the experimental sharding
mode we must declare the full shape on every jax process, therefore this returns
:attr:`~Sampler.n_chains`.
Usage of this flag is required to support both MPI and sharding.
Samplers may override this to have a larger batch size, for example to
propagate multiple replicas (in the case of parallel tempering).
"""
if config.netket_experimental_sharding:
return self.n_chains
else:
return self.n_chains_per_rank
@property
def is_exact(self) -> bool:
"""
Returns `True` if the sampler is exact.
The sampler is exact if all the samples are exactly distributed according to the
chosen power of the variational state, and there is no correlation among them.
"""
return False
[docs]
def log_pdf(self, model: Union[Callable, nn.Module]) -> Callable:
"""
Returns a closure with the log-pdf function encoded by this sampler.
Args:
model: A Flax module or callable with the forward pass of the log-pdf.
If it is a callable, it should have the signature :code:`f(parameters, σ) -> jnp.ndarray`.
Returns:
The log-probability density function.
Note:
The result is returned as a `HashablePartial` so that the closure
does not trigger recompilation.
"""
apply_fun = get_afun_if_module(model)
log_pdf = HashablePartial(
lambda apply_fun, pars, σ: self.machine_pow * apply_fun(pars, σ).real,
apply_fun,
)
return log_pdf
[docs]
def init_state(
sampler,
machine: Union[Callable, nn.Module],
parameters: PyTree,
seed: Optional[SeedT] = None,
) -> SamplerState:
"""
Creates the structure holding the state of the sampler.
If you want reproducible samples, you should specify `seed`, otherwise the state
will be initialised randomly.
If running across several MPI processes, all `sampler_state`s are guaranteed to be
in a different (but deterministic) state.
This is achieved by first reducing (summing) the seed provided to every MPI rank,
then generating `n_rank` seeds starting from the reduced one, and every rank is
initialized with one of those seeds.
The resulting state is guaranteed to be a frozen Python dataclass (in particular,
a Flax dataclass), and it can be serialized using Flax serialization methods.
Args:
machine: A Flax module or callable with the forward pass of the log-pdf.
If it is a callable, it should have the signature :code:`f(parameters, σ) -> jnp.ndarray`.
parameters: The PyTree of parameters of the model.
seed: An optional seed or jax PRNGKey. If not specified, a random seed will be used.
Returns:
The structure holding the state of the sampler. In general you should not expect
it to be in a valid state, and should reset it before use.
"""
key = nkjax.PRNGKey(seed)
key = nkjax.mpi_split(key)
return sampler._init_state(wrap_afun(machine), parameters, key)
[docs]
def reset(
sampler,
machine: Union[Callable, nn.Module],
parameters: PyTree,
state: Optional[SamplerState] = None,
) -> SamplerState:
"""
Resets the state of the sampler. To be used every time the parameters are changed.
Args:
machine: A Flax module or callable with the forward pass of the log-pdf.
If it is a callable, it should have the signature :code:`f(parameters, σ) -> jnp.ndarray`.
parameters: The PyTree of parameters of the model.
state: The current state of the sampler. If not specified, it will be constructed
by calling :code:`sampler.init_state(machine, parameters)` with a random seed.
Returns:
A valid sampler state.
"""
if state is None:
state = sampler.init_state(machine, parameters)
return sampler._reset(wrap_afun(machine), parameters, state)
[docs]
def sample(
sampler,
machine: Union[Callable, nn.Module],
parameters: PyTree,
*,
state: Optional[SamplerState] = None,
chain_length: int = 1,
) -> tuple[jnp.ndarray, SamplerState]:
"""
Samples `chain_length` batches of samples along the chains.
Arguments:
machine: A Flax module or callable with the forward pass of the log-pdf.
If it is a callable, it should have the signature :code:`f(parameters, σ) -> jnp.ndarray`.
parameters: The PyTree of parameters of the model.
state: The current state of the sampler. If not specified, then initialize and reset it.
chain_length: The length of the chains (default = 1).
Returns:
σ: The generated batches of samples.
state: The new state of the sampler.
"""
if state is None:
state = sampler.reset(machine, parameters)
return sampler._sample_chain(
wrap_afun(machine), parameters, state, chain_length
)
[docs]
def samples(
sampler,
machine: Union[Callable, nn.Module],
parameters: PyTree,
*,
state: Optional[SamplerState] = None,
chain_length: int = 1,
) -> Iterator[jnp.ndarray]:
"""
Returns a generator sampling `chain_length` batches of samples along the chains.
Arguments:
machine: A Flax module or callable with the forward pass of the log-pdf.
If it is a callable, it should have the signature :code:`f(parameters, σ) -> jnp.ndarray`.
parameters: The PyTree of parameters of the model.
state: The current state of the sampler. If not specified, then initialize and reset it.
chain_length: The length of the chains (default = 1).
"""
if state is None:
state = sampler.reset(machine, parameters)
machine = wrap_afun(machine)
for _i in range(chain_length):
samples, state = sampler._sample_chain(machine, parameters, state, 1)
yield samples[:, 0, :]
@abc.abstractmethod
def _sample_chain(
sampler,
machine: nn.Module,
parameters: PyTree,
state: SamplerState,
chain_length: int,
) -> tuple[jnp.ndarray, SamplerState]:
"""
Implementation of `sample` for subclasses of `Sampler`.
If you subclass `Sampler`, you should override this and not `sample`
itself, because `sample` contains some common logic.
If using Jax, this function should be jitted.
Arguments:
machine: A Flax module with the forward pass of the log-pdf.
parameters: The PyTree of parameters of the model.
state: The current state of the sampler.
chain_length: The length of the chains.
Returns:
σ: The generated batches of samples.
state: The new state of the sampler.
"""
@abc.abstractmethod
def _init_state(sampler, machine, params, seed) -> SamplerState:
"""
Implementation of `init_state` for subclasses of `Sampler`.
If you subclass `Sampler`, you should override this and not `init_state`
itself, because `init_state` contains some common logic.
"""
@abc.abstractmethod
def _reset(sampler, machine, parameters, state):
"""
Implementation of `reset` for subclasses of `Sampler`.
If you subclass `Sampler`, you should override this and not `reset`
itself, because `reset` contains some common logic.
"""
