#!/usr/bin/env python3
# Copyright (c) Meta Platforms, Inc. and affiliates.
#
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
from __future__ import annotations
from copy import deepcopy
from typing import Any, Callable, Dict, List, Optional, Tuple
import numpy as np
import torch
from ax.core.search_space import SearchSpaceDigest
from ax.core.types import TCandidateMetadata, TGenMetadata
from ax.models.torch.botorch_defaults import (
get_and_fit_model,
get_NEI,
recommend_best_observed_point,
scipy_optimizer,
)
from ax.models.torch.utils import (
_get_X_pending_and_observed,
_to_inequality_constraints,
normalize_indices,
predict_from_model,
subset_model,
)
from ax.models.torch_base import TorchModel
from ax.models.types import TConfig
from ax.utils.common.constants import Keys
from ax.utils.common.docutils import copy_doc
from ax.utils.common.logger import get_logger
from ax.utils.common.typeutils import checked_cast
from botorch.acquisition.acquisition import AcquisitionFunction
from botorch.models.model import Model
from botorch.models.model_list_gp_regression import ModelListGP
from torch import Tensor
logger = get_logger(__name__)
TModelConstructor = Callable[
[
List[Tensor],
List[Tensor],
List[Tensor],
List[int],
List[int],
List[str],
Optional[Dict[str, Tensor]],
Any,
],
Model,
]
TModelPredictor = Callable[[Model, Tensor], Tuple[Tensor, Tensor]]
TAcqfConstructor = Callable[
[
Model,
Tensor,
Optional[Tuple[Tensor, Tensor]],
Optional[Tensor],
Optional[Tensor],
Any,
],
AcquisitionFunction,
]
TOptimizer = Callable[
[
AcquisitionFunction,
Tensor,
int,
Optional[List[Tuple[Tensor, Tensor, float]]],
Optional[List[Tuple[Tensor, Tensor, float]]],
Optional[Dict[int, float]],
Optional[Callable[[Tensor], Tensor]],
Any,
],
Tuple[Tensor, Tensor],
]
TBestPointRecommender = Callable[
[
TorchModel,
List[Tuple[float, float]],
Tensor,
Optional[Tuple[Tensor, Tensor]],
Optional[Tuple[Tensor, Tensor]],
Optional[Dict[int, float]],
Optional[TConfig],
Optional[Dict[int, float]],
],
Optional[Tensor],
]
[docs]class BotorchModel(TorchModel):
r"""
Customizable botorch model.
By default, this uses a noisy Expected Improvement acquisition function on
top of a model made up of separate GPs, one for each outcome. This behavior
can be modified by providing custom implementations of the following
components:
- a `model_constructor` that instantiates and fits a model on data
- a `model_predictor` that predicts outcomes using the fitted model
- a `acqf_constructor` that creates an acquisition function from a fitted model
- a `acqf_optimizer` that optimizes the acquisition function
- a `best_point_recommender` that recommends a current "best" point (i.e.,
what the model recommends if the learning process ended now)
Args:
model_constructor: A callable that instantiates and fits a model on data,
with signature as described below.
model_predictor: A callable that predicts using the fitted model, with
signature as described below.
acqf_constructor: A callable that creates an acquisition function from a
fitted model, with signature as described below.
acqf_optimizer: A callable that optimizes the acquisition function, with
signature as described below.
best_point_recommender: A callable that recommends the best point, with
signature as described below.
refit_on_cv: If True, refit the model for each fold when performing
cross-validation.
refit_on_update: If True, refit the model after updating the training
data using the `update` method.
warm_start_refitting: If True, start model refitting from previous
model parameters in order to speed up the fitting process.
Call signatures:
::
model_constructor(
Xs,
Ys,
Yvars,
task_features,
fidelity_features,
metric_names,
state_dict,
**kwargs,
) -> model
Here `Xs`, `Ys`, `Yvars` are lists of tensors (one element per outcome),
`task_features` identifies columns of Xs that should be modeled as a task,
`fidelity_features` is a list of ints that specify the positions of fidelity
parameters in 'Xs', `metric_names` provides the names of each `Y` in `Ys`,
`state_dict` is a pytorch module state dict, and `model` is a BoTorch `Model`.
Optional kwargs are being passed through from the `BotorchModel` constructor.
This callable is assumed to return a fitted BoTorch model that has the same
dtype and lives on the same device as the input tensors.
::
model_predictor(model, X) -> [mean, cov]
Here `model` is a fitted botorch model, `X` is a tensor of candidate points,
and `mean` and `cov` are the posterior mean and covariance, respectively.
::
acqf_constructor(
model,
objective_weights,
outcome_constraints,
X_observed,
X_pending,
**kwargs,
) -> acq_function
Here `model` is a botorch `Model`, `objective_weights` is a tensor of weights
for the model outputs, `outcome_constraints` is a tuple of tensors describing
the (linear) outcome constraints, `X_observed` are previously observed points,
and `X_pending` are points whose evaluation is pending. `acq_function` is a
BoTorch acquisition function crafted from these inputs. For additional
details on the arguments, see `get_NEI`.
::
acqf_optimizer(
acq_function,
bounds,
n,
inequality_constraints,
equality_constraints,
fixed_features,
rounding_func,
**kwargs,
) -> candidates
Here `acq_function` is a BoTorch `AcquisitionFunction`, `bounds` is a tensor
containing bounds on the parameters, `n` is the number of candidates to be
generated, `inequality_constraints` are inequality constraints on parameter
values, `fixed_features` specifies features that should be fixed during
generation, and `rounding_func` is a callback that rounds an optimization
result appropriately. `candidates` is a tensor of generated candidates.
For additional details on the arguments, see `scipy_optimizer`.
::
best_point_recommender(
model,
bounds,
objective_weights,
outcome_constraints,
linear_constraints,
fixed_features,
model_gen_options,
target_fidelities,
) -> candidates
Here `model` is a TorchModel, `bounds` is a list of tuples containing bounds
on the parameters, `objective_weights` is a tensor of weights for the model outputs,
`outcome_constraints` is a tuple of tensors describing the (linear) outcome
constraints, `linear_constraints` is a tuple of tensors describing constraints
on the design, `fixed_features` specifies features that should be fixed during
generation, `model_gen_options` is a config dictionary that can contain
model-specific options, and `target_fidelities` is a map from fidelity feature
column indices to their respective target fidelities, used for multi-fidelity
optimization problems. % TODO: refer to an example.
"""
dtype: Optional[torch.dtype]
device: Optional[torch.device]
Xs: List[Tensor]
Ys: List[Tensor]
Yvars: List[Tensor]
model: Optional[Model]
def __init__(
self,
model_constructor: TModelConstructor = get_and_fit_model,
model_predictor: TModelPredictor = predict_from_model,
# pyre-fixme[9]: acqf_constructor has type `Callable[[Model, Tensor,
# Optional[Tuple[Tensor, Tensor]], Optional[Tensor], Optional[Tensor], Any],
# AcquisitionFunction]`; used as `Callable[[Model, Tensor,
# Optional[Tuple[Tensor, Tensor]], Optional[Tensor], Optional[Tensor],
# **(Any)], AcquisitionFunction]`.
acqf_constructor: TAcqfConstructor = get_NEI,
# pyre-fixme[9]: acqf_optimizer declared/used type mismatch
acqf_optimizer: TOptimizer = scipy_optimizer,
best_point_recommender: TBestPointRecommender = recommend_best_observed_point,
refit_on_cv: bool = False,
refit_on_update: bool = True,
warm_start_refitting: bool = True,
use_input_warping: bool = False,
use_loocv_pseudo_likelihood: bool = False,
**kwargs: Any,
) -> None:
self.model_constructor = model_constructor
self.model_predictor = model_predictor
self.acqf_constructor = acqf_constructor
self.acqf_optimizer = acqf_optimizer
self.best_point_recommender = best_point_recommender
self._kwargs = kwargs
self.refit_on_cv = refit_on_cv
self.refit_on_update = refit_on_update
self.warm_start_refitting = warm_start_refitting
self.use_input_warping = use_input_warping
self.use_loocv_pseudo_likelihood = use_loocv_pseudo_likelihood
self.model: Optional[Model] = None
self.Xs = []
self.Ys = []
self.Yvars = []
self.dtype = None
self.device = None
self.task_features: List[int] = []
self.fidelity_features: List[int] = []
self.metric_names: List[str] = []
[docs] @copy_doc(TorchModel.fit)
def fit(
self,
Xs: List[Tensor],
Ys: List[Tensor],
Yvars: List[Tensor],
search_space_digest: SearchSpaceDigest,
metric_names: List[str],
candidate_metadata: Optional[List[List[TCandidateMetadata]]] = None,
) -> None:
self.dtype = Xs[0].dtype
self.device = Xs[0].device
self.Xs = Xs
self.Ys = Ys
self.Yvars = Yvars
self.task_features = normalize_indices(
search_space_digest.task_features, d=Xs[0].size(-1)
)
self.fidelity_features = normalize_indices(
search_space_digest.fidelity_features, d=Xs[0].size(-1)
)
self.metric_names = metric_names
self.model = self.model_constructor( # pyre-ignore [28]
Xs=Xs,
Ys=Ys,
Yvars=Yvars,
task_features=self.task_features,
fidelity_features=self.fidelity_features,
metric_names=self.metric_names,
use_input_warping=self.use_input_warping,
use_loocv_pseudo_likelihood=self.use_loocv_pseudo_likelihood,
**self._kwargs,
)
[docs] @copy_doc(TorchModel.predict)
def predict(self, X: Tensor) -> Tuple[Tensor, Tensor]:
return self.model_predictor(model=self.model, X=X) # pyre-ignore [28]
[docs] @copy_doc(TorchModel.gen)
def gen(
self,
n: int,
bounds: List[Tuple[float, float]],
objective_weights: Tensor,
outcome_constraints: Optional[Tuple[Tensor, Tensor]] = None,
linear_constraints: Optional[Tuple[Tensor, Tensor]] = None,
fixed_features: Optional[Dict[int, float]] = None,
pending_observations: Optional[List[Tensor]] = None,
model_gen_options: Optional[TConfig] = None,
rounding_func: Optional[Callable[[Tensor], Tensor]] = None,
target_fidelities: Optional[Dict[int, float]] = None,
) -> Tuple[Tensor, Tensor, TGenMetadata, Optional[List[TCandidateMetadata]]]:
options = model_gen_options or {}
acf_options = options.get(Keys.ACQF_KWARGS, {})
optimizer_options = options.get(Keys.OPTIMIZER_KWARGS, {})
if target_fidelities:
raise NotImplementedError(
"target_fidelities not implemented for base BotorchModel"
)
X_pending, X_observed = _get_X_pending_and_observed(
Xs=self.Xs,
pending_observations=pending_observations,
objective_weights=objective_weights,
outcome_constraints=outcome_constraints,
bounds=bounds,
linear_constraints=linear_constraints,
fixed_features=fixed_features,
)
model = self.model
# subset model only to the outcomes we need for the optimization 357
if options.get(Keys.SUBSET_MODEL, True):
subset_model_results = subset_model(
model=model, # pyre-ignore [6]
objective_weights=objective_weights,
outcome_constraints=outcome_constraints,
)
model = subset_model_results.model
objective_weights = subset_model_results.objective_weights
outcome_constraints = subset_model_results.outcome_constraints
bounds_ = torch.tensor(bounds, dtype=self.dtype, device=self.device)
bounds_ = bounds_.transpose(0, 1)
botorch_rounding_func = get_rounding_func(rounding_func)
# The following logic is to work around the limitation of PyTorch's Sobol
# sampler to <1111 dimensions.
# TODO: Remove once https://github.com/pytorch/pytorch/issues/41489 is resolved.
from botorch.exceptions.errors import UnsupportedError
def make_and_optimize_acqf(override_qmc: bool = False) -> Tuple[Tensor, Tensor]:
add_kwargs = {"qmc": False} if override_qmc else {}
acquisition_function = self.acqf_constructor( # pyre-ignore: [28]
model=model,
objective_weights=objective_weights,
outcome_constraints=outcome_constraints,
X_observed=X_observed,
X_pending=X_pending,
**acf_options,
**add_kwargs,
)
acquisition_function = checked_cast(
AcquisitionFunction, acquisition_function
)
# pyre-ignore: [28]
candidates, expected_acquisition_value = self.acqf_optimizer(
acq_function=checked_cast(AcquisitionFunction, acquisition_function),
bounds=bounds_,
n=n,
inequality_constraints=_to_inequality_constraints(
linear_constraints=linear_constraints
),
fixed_features=fixed_features,
rounding_func=botorch_rounding_func,
**optimizer_options,
)
return candidates, expected_acquisition_value
try:
candidates, expected_acquisition_value = make_and_optimize_acqf()
except UnsupportedError as e:
if "SobolQMCSampler only supports dimensions q * o <= 1111" in str(e):
# dimension too large for Sobol, let's use IID
candidates, expected_acquisition_value = make_and_optimize_acqf(
override_qmc=True
)
else:
raise e
return (
candidates.detach().cpu(),
torch.ones(n, dtype=self.dtype),
{"expected_acquisition_value": expected_acquisition_value.tolist()},
None,
)
[docs] @copy_doc(TorchModel.best_point)
def best_point(
self,
bounds: List[Tuple[float, float]],
objective_weights: Tensor,
outcome_constraints: Optional[Tuple[Tensor, Tensor]] = None,
linear_constraints: Optional[Tuple[Tensor, Tensor]] = None,
fixed_features: Optional[Dict[int, float]] = None,
model_gen_options: Optional[TConfig] = None,
target_fidelities: Optional[Dict[int, float]] = None,
) -> Optional[Tensor]:
return self.best_point_recommender( # pyre-ignore [28]
model=self,
bounds=bounds,
objective_weights=objective_weights,
outcome_constraints=outcome_constraints,
linear_constraints=linear_constraints,
fixed_features=fixed_features,
model_gen_options=model_gen_options,
target_fidelities=target_fidelities,
)
[docs] @copy_doc(TorchModel.cross_validate)
def cross_validate( # pyre-ignore[14]: Some `TorchModel.cross_validate` kwargs
self, # are not needed here and therefore we just use `**kwargs` catchall.
Xs_train: List[Tensor],
Ys_train: List[Tensor],
Yvars_train: List[Tensor],
X_test: Tensor,
**kwargs: Any,
) -> Tuple[Tensor, Tensor]:
if self.model is None:
raise RuntimeError("Cannot cross-validate model that has not been fitted")
if self.refit_on_cv:
state_dict = None
else:
state_dict = deepcopy(self.model.state_dict())
model = self.model_constructor( # pyre-ignore: [28]
Xs=Xs_train,
Ys=Ys_train,
Yvars=Yvars_train,
task_features=self.task_features,
state_dict=state_dict,
fidelity_features=self.fidelity_features,
metric_names=self.metric_names,
refit_model=self.refit_on_cv,
use_input_warping=self.use_input_warping,
use_loocv_pseudo_likelihood=self.use_loocv_pseudo_likelihood,
**self._kwargs,
)
return self.model_predictor(model=model, X=X_test) # pyre-ignore: [28]
[docs] @copy_doc(TorchModel.update)
def update( # pyre-ignore[14]: Some `TorchModel.update` kwargs are not
self, # needed here and therefore we just use `**kwargs` catchall.
Xs: List[Tensor],
Ys: List[Tensor],
Yvars: List[Tensor],
candidate_metadata: Optional[List[List[TCandidateMetadata]]] = None,
**kwargs: Any,
) -> None:
if self.model is None:
raise RuntimeError("Cannot update model that has not been fitted")
self.Xs = Xs
self.Ys = Ys
self.Yvars = Yvars
if self.refit_on_update and not self.warm_start_refitting:
state_dict = None # pragma: no cover
else:
state_dict = deepcopy(self.model.state_dict())
self.model = self.model_constructor( # pyre-ignore: [28]
Xs=self.Xs,
Ys=self.Ys,
Yvars=self.Yvars,
task_features=self.task_features,
state_dict=state_dict,
fidelity_features=self.fidelity_features,
metric_names=self.metric_names,
refit_model=self.refit_on_update,
use_input_warping=self.use_input_warping,
use_loocv_pseudo_likelihood=self.use_loocv_pseudo_likelihood,
**self._kwargs,
)
[docs] def feature_importances(self) -> np.ndarray:
if self.model is None:
raise RuntimeError(
"Cannot calculate feature_importances without a fitted model"
)
elif isinstance(self.model, ModelListGP):
models = self.model.models
else:
models = [self.model]
lengthscales = []
for m in models:
ls = m.covar_module.base_kernel.lengthscale
if ls.ndim == 2:
ls = ls.unsqueeze(0)
lengthscales.append(ls)
lengthscales = torch.cat(lengthscales, dim=0)
return (1 / lengthscales).detach().cpu().numpy()
[docs]def get_rounding_func(
rounding_func: Optional[Callable[[Tensor], Tensor]]
) -> Optional[Callable[[Tensor], Tensor]]:
if rounding_func is None:
botorch_rounding_func = rounding_func
else:
# make sure rounding_func is properly applied to q- and t-batches
def botorch_rounding_func(X: Tensor) -> Tensor:
batch_shape, d = X.shape[:-1], X.shape[-1]
X_round = torch.stack(
[rounding_func(x) for x in X.view(-1, d)] # pyre-ignore: [16]
)
return X_round.view(*batch_shape, d)
return botorch_rounding_func
