index int64 0 50.7k | project_path stringclasses 279
values | completion_path stringlengths 57 199 | signature_postion stringlengths 3 11 | class_position stringlengths 3 11 | function_name stringlengths 1 62 | class_name stringlengths 0 55 | body_position stringlengths 3 11 | body_position_with_comments stringlengths 3 11 | masked_class stringlengths 14 8.32k | masked_class_with_comment stringlengths 14 8.32k | func_body stringlengths 9 6.72k | func_body_with_comment stringlengths 9 6.72k | len_func_body int64 1 643 | len_input int64 10 2.02k | len_output int64 4 1.9k | total int64 20 2.05k | file_code stringlengths 14 442k | file_code_with_comment stringlengths 14 442k |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
46,929 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/designers/random_test.py | 48,49 | 30,103 | test_log_scaling | RandomTest | 50,82 | 49,82 | class RandomTest(absltest.TestCase):
def test_on_flat_space(self):
config = vz.ProblemStatement(test_studies.flat_space_with_all_types())
designer = random.RandomDesigner(config.search_space, seed=None)
self.assertLen(
test_runners.run_with_random_metrics(
designer, config, iters=50, ... | class RandomTest(absltest.TestCase):
def test_on_flat_space(self):
config = vz.ProblemStatement(test_studies.flat_space_with_all_types())
designer = random.RandomDesigner(config.search_space, seed=None)
self.assertLen(
test_runners.run_with_random_metrics(
designer, config, iters=50, ... | config = vz.ProblemStatement(test_studies.flat_space_with_all_types())
designer = random.RandomDesigner(config.search_space, seed=None)
trials = test_runners.run_with_random_metrics(
designer, config, iters=1000, batch_size=1
)
# LINEAR scaling.
sum_lineardouble = 0.0
for t in trials... | """Confirm that LINEAR and LOG scaling give the right distribution."""
config = vz.ProblemStatement(test_studies.flat_space_with_all_types())
designer = random.RandomDesigner(config.search_space, seed=None)
trials = test_runners.run_with_random_metrics(
designer, config, iters=1000, batch_size=1... | 171 | 571 | 564 | 1,135 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,930 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/designers/scheduled_designer.py | 90,91 | 84,98 | value | ExponentialScheduledParam | 91,98 | 91,98 | @attrs.define
class ExponentialScheduledParam(ScheduledParam):
"""Exponential scheduler."""
rate: float
def value(self, total_steps: int, step: int) -> float:
FILL_FUNC_BODY | @attrs.define
class ExponentialScheduledParam(ScheduledParam):
"""Exponential scheduler."""
rate: float
def value(self, total_steps: int, step: int) -> float:
FILL_FUNC_BODY | alpha = 1 / self.rate
beta = -np.log(self.final_value / self.init_value) / (
(total_steps - 1) ** alpha
)
return self.init_value * np.exp(-beta * step**alpha)
| alpha = 1 / self.rate
beta = -np.log(self.final_value / self.init_value) / (
(total_steps - 1) ** alpha
)
return self.init_value * np.exp(-beta * step**alpha)
| 24 | 64 | 75 | 139 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,931 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/designers/scheduled_designer.py | 79,80 | 75,84 | value | LinearScheduledParam | 80,84 | 80,84 | @attrs.define
class LinearScheduledParam(ScheduledParam):
"""Linear scheduler."""
def value(self, total_steps: int, step: int) -> float:
FILL_FUNC_BODY | @attrs.define
class LinearScheduledParam(ScheduledParam):
"""Linear scheduler."""
def value(self, total_steps: int, step: int) -> float:
FILL_FUNC_BODY | slope = (self.final_value - self.init_value) / (total_steps - 1)
return self.init_value + slope * step
| slope = (self.final_value - self.init_value) / (total_steps - 1)
return self.init_value + slope * step
| 15 | 57 | 44 | 101 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,932 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/designers/scheduled_designer.py | 197,200 | 118,254 | _validate_num_incorporated_suggested_trials | ScheduledDesigner | 201,215 | 200,215 | @attrs.define
class ScheduledDesigner(vza.PartiallySerializableDesigner):
"""Scheduled designer."""
_problem: vz.ProblemStatement = attrs.field(kw_only=False)
_designer_factory: vza.DesignerFactory = attrs.field(kw_only=True)
_designer_state_updater: DesignerStateUpdater = attrs.field(kw_only=True)
_schedule... | @attrs.define
class ScheduledDesigner(vza.PartiallySerializableDesigner):
"""Scheduled designer."""
_problem: vz.ProblemStatement = attrs.field(kw_only=False)
_designer_factory: vza.DesignerFactory = attrs.field(kw_only=True)
_designer_state_updater: DesignerStateUpdater = attrs.field(kw_only=True)
_schedule... | num_completed = len(completed.trials)
num_active = len(all_active.trials)
if self._num_incorporated_suggested_trials < num_completed + num_active:
# If the designer is updated with more trials than what it observed, we
# fast forward and update the state. Would happen if the designer is
# ... | """Validate (and if needed update) the designer's suggested number of trials."""
num_completed = len(completed.trials)
num_active = len(all_active.trials)
if self._num_incorporated_suggested_trials < num_completed + num_active:
# If the designer is updated with more trials than what it observed, w... | 74 | 1,515 | 225 | 1,740 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,933 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/designers/scheduled_designer.py | 52,53 | 33,75 | _validate | ScheduledParam | 54,63 | 53,63 | @attrs.define
class ScheduledParam(abc.ABC):
"""Represents a scheduled parameter.
The scheduling starts from `init_value` (the first step) and ends at
`final_value` (last step).
The `value` method returns the value of the scheduled param in the current
step, accounting for the total number of steps in the s... | @attrs.define
class ScheduledParam(abc.ABC):
"""Represents a scheduled parameter.
The scheduling starts from `init_value` (the first step) and ends at
`final_value` (last step).
The `value` method returns the value of the scheduled param in the current
step, accounting for the total number of steps in the s... | total_steps = 10 # arbitrarily choose the total steps for validation.
tol = 1e-5
if abs(self.value(total_steps, 0) - self.init_value) > tol:
raise ValueError("Initial value mismatch (step 0).")
if abs(self.value(total_steps, total_steps - 1) - self.final_value) > tol:
raise ValueError("Fi... | """Validate the scheduling param initial and final values."""
total_steps = 10 # arbitrarily choose the total steps for validation.
tol = 1e-5
if abs(self.value(total_steps, 0) - self.init_value) > tol:
raise ValueError("Initial value mismatch (step 0).")
if abs(self.value(total_steps, tota... | 42 | 294 | 130 | 424 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,934 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/designers/scheduled_designer_test.py | 49,50 | 34,54 | suggest | MockParameterizedDesigner | 51,54 | 50,54 | @attrs.define
class MockParameterizedDesigner(vza.Designer):
"""Mock parameterized designer."""
problem: vz.ProblemStatement
parameter1: float = 0.0
parameter2: float = 0.0
def update(
self, completed: vza.CompletedTrials, all_active: vza.ActiveTrials
) -> None:
"""Update the designer based on c... | @attrs.define
class MockParameterizedDesigner(vza.Designer):
"""Mock parameterized designer."""
problem: vz.ProblemStatement
parameter1: float = 0.0
parameter2: float = 0.0
def update(
self, completed: vza.CompletedTrials, all_active: vza.ActiveTrials
) -> None:
"""Update the designer based on c... | return random.RandomDesigner(self.problem.search_space).suggest(count)
| """Suggest trials."""
return random.RandomDesigner(self.problem.search_space).suggest(count)
| 2 | 164 | 26 | 190 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,935 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/designers/scheduled_designer_test.py | 149,150 | 65,191 | test_scheduled_designer_serialization | ScheduledDesignerTest | 150,191 | 150,191 | class ScheduledDesignerTest(absltest.TestCase):
def test_schedule_designer(self):
expected_total_num_trials = 10
problem = vz.ProblemStatement(
test_studies.flat_continuous_space_with_scaling()
)
problem.metric_information.append(
vz.MetricInformation(
name="metric", goal=... | class ScheduledDesignerTest(absltest.TestCase):
def test_schedule_designer(self):
expected_total_num_trials = 10
problem = vz.ProblemStatement(
test_studies.flat_continuous_space_with_scaling()
)
problem.metric_information.append(
vz.MetricInformation(
name="metric", goal=... | expected_total_num_trials = 10
problem = vz.ProblemStatement(
test_studies.flat_continuous_space_with_scaling()
)
problem.metric_information.append(
vz.MetricInformation(
name="metric", goal=vz.ObjectiveMetricGoal.MAXIMIZE
)
)
param1 = scheduled_designer.Linea... | expected_total_num_trials = 10
problem = vz.ProblemStatement(
test_studies.flat_continuous_space_with_scaling()
)
problem.metric_information.append(
vz.MetricInformation(
name="metric", goal=vz.ObjectiveMetricGoal.MAXIMIZE
)
)
param1 = scheduled_designer.Linea... | 84 | 1,091 | 542 | 1,633 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,936 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/designers/scheduled_designer_test.py | 193,194 | 191,227 | test_scheduled_gp_bandit | ScheduledGpBanditTest | 194,227 | 194,227 | class ScheduledGpBanditTest(absltest.TestCase):
def test_scheduled_gp_bandit(self):
FILL_FUNC_BODY | class ScheduledGpBanditTest(absltest.TestCase):
def test_scheduled_gp_bandit(self):
FILL_FUNC_BODY | problem = vz.ProblemStatement(
test_studies.flat_continuous_space_with_scaling()
)
problem.metric_information.append(
vz.MetricInformation(
name="metric", goal=vz.ObjectiveMetricGoal.MAXIMIZE
)
)
def _gp_bandit_factory(problem):
return gp_bandit.VizierGPBan... | problem = vz.ProblemStatement(
test_studies.flat_continuous_space_with_scaling()
)
problem.metric_information.append(
vz.MetricInformation(
name="metric", goal=vz.ObjectiveMetricGoal.MAXIMIZE
)
)
def _gp_bandit_factory(problem):
return gp_bandit.VizierGPBan... | 35 | 44 | 331 | 375 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,937 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/designers/scheduled_designer_test.py | 229,230 | 227,271 | test_scheduled_gp_ucb_pe | ScheduledGpUcbPeTest | 230,271 | 230,271 | class ScheduledGpUcbPeTest(absltest.TestCase):
def test_scheduled_gp_ucb_pe(self):
FILL_FUNC_BODY | class ScheduledGpUcbPeTest(absltest.TestCase):
def test_scheduled_gp_ucb_pe(self):
FILL_FUNC_BODY | problem = vz.ProblemStatement(
test_studies.flat_continuous_space_with_scaling()
)
problem.metric_information.append(
vz.MetricInformation(
name="metric", goal=vz.ObjectiveMetricGoal.MAXIMIZE
)
)
def _gp_ucb_pe_factory(
problem: vz.ProblemStatement,
)... | problem = vz.ProblemStatement(
test_studies.flat_continuous_space_with_scaling()
)
problem.metric_information.append(
vz.MetricInformation(
name="metric", goal=vz.ObjectiveMetricGoal.MAXIMIZE
)
)
def _gp_ucb_pe_factory(
problem: vz.ProblemStatement,
)... | 46 | 46 | 494 | 540 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,938 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/designers/scheduled_gp_bandit.py | 44,45 | 28,64 | _gp_bandit_state_updater | ScheduledGPBanditFactory | 45,51 | 45,51 | @attrs.define(kw_only=True)
class ScheduledGPBanditFactory:
"""Scheduled GP-Bandit factory."""
_gp_bandit_factory: Callable[[vz.ProblemStatement], gp_bandit.VizierGPBandit]
_init_ucb_coefficient: float
_final_ucb_coefficient: float
_decay_ucb_coefficient: float
_expected_total_num_trials: int
def __call... | @attrs.define(kw_only=True)
class ScheduledGPBanditFactory:
"""Scheduled GP-Bandit factory."""
_gp_bandit_factory: Callable[[vz.ProblemStatement], gp_bandit.VizierGPBandit]
_init_ucb_coefficient: float
_final_ucb_coefficient: float
_decay_ucb_coefficient: float
_expected_total_num_trials: int
def __call... | designer._scoring_function_factory = ( # pylint: disable=protected-access
acquisitions.bayesian_scoring_function_factory(
lambda _: acquisitions.UCB(params['ucb_coefficient'])
)
)
| designer._scoring_function_factory = ( # pylint: disable=protected-access
acquisitions.bayesian_scoring_function_factory(
lambda _: acquisitions.UCB(params['ucb_coefficient'])
)
)
| 12 | 377 | 71 | 448 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,939 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/designers/scheduled_gp_ucb_pe.py | 57,58 | 28,107 | _gp_ucb_pe_state_updater | ScheduledGPUCBPEFactory | 58,72 | 58,72 | @attrs.define(kw_only=True)
class ScheduledGPUCBPEFactory:
"""Scheduled GP-UCB-PE factory."""
_gp_ucb_pe_factory: Callable[
[vz.ProblemStatement], gp_ucb_pe.VizierGPUCBPEBandit
]
_expected_total_num_trials: int
# ucb_coefficient
_init_ucb_coefficient: float
_final_ucb_coefficient: float
_decay_uc... | @attrs.define(kw_only=True)
class ScheduledGPUCBPEFactory:
"""Scheduled GP-UCB-PE factory."""
_gp_ucb_pe_factory: Callable[
[vz.ProblemStatement], gp_ucb_pe.VizierGPUCBPEBandit
]
_expected_total_num_trials: int
# ucb_coefficient
_init_ucb_coefficient: float
_final_ucb_coefficient: float
_decay_uc... | # Create a new copy of the config and replace the ucb coefficient value.
updated_config = dc.replace(
designer._config, # pylint: disable=protected-access
ucb_coefficient=params['ucb_coefficient'],
explore_region_ucb_coefficient=params[
'explore_region_ucb_coeffi... | # Create a new copy of the config and replace the ucb coefficient value.
updated_config = dc.replace(
designer._config, # pylint: disable=protected-access
ucb_coefficient=params['ucb_coefficient'],
explore_region_ucb_coefficient=params[
'explore_region_ucb_coeffi... | 67 | 785 | 210 | 995 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,940 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/designers/unsafe_as_infeasible_designer.py | 55,58 | 27,93 | update | UnsafeAsInfeasibleDesigner | 58,80 | 58,80 | class UnsafeAsInfeasibleDesigner(vza.Designer):
"""Designer that maps unsafe Trials as infeasible."""
def __init__(
self,
problem_statement: vz.ProblemStatement,
designer_factory: vza.DesignerFactory[vza.Designer],
*,
seed: Optional[int] = None,
):
"""Init.
Args:
prob... | class UnsafeAsInfeasibleDesigner(vza.Designer):
"""Designer that maps unsafe Trials as infeasible."""
def __init__(
self,
problem_statement: vz.ProblemStatement,
designer_factory: vza.DesignerFactory[vza.Designer],
*,
seed: Optional[int] = None,
):
"""Init.
Args:
prob... | are_safe = self._safety_checker.are_trials_safe(completed.trials)
mutated_trials = []
mutated_trial_count = 0
for is_safe, trial in zip(are_safe, completed.trials):
if is_safe:
mutated_trials.append(trial)
else:
measurement = trial.final_measurement or vz.Measurement()
... | are_safe = self._safety_checker.are_trials_safe(completed.trials)
mutated_trials = []
mutated_trial_count = 0
for is_safe, trial in zip(are_safe, completed.trials):
if is_safe:
mutated_trials.append(trial)
else:
measurement = trial.final_measurement or vz.Measurement()
... | 54 | 404 | 275 | 679 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,941 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/designers/unsafe_as_infeasible_designer_test.py | 64,65 | 27,112 | test_unsafe_multi_objective_eagle | UnsafeAsInfeasibleDesignerTest | 65,112 | 65,112 | class UnsafeAsInfeasibleDesignerTest(absltest.TestCase):
def test_unsafe_eagle(self):
problem = vz.ProblemStatement(
test_studies.flat_continuous_space_with_scaling()
)
problem.metric_information.extend([
vz.MetricInformation(
name='metric1', goal=vz.ObjectiveMetricGoal.MAXIMI... | class UnsafeAsInfeasibleDesignerTest(absltest.TestCase):
def test_unsafe_eagle(self):
problem = vz.ProblemStatement(
test_studies.flat_continuous_space_with_scaling()
)
problem.metric_information.extend([
vz.MetricInformation(
name='metric1', goal=vz.ObjectiveMetricGoal.MAXIMI... | problem = vz.ProblemStatement(
test_studies.flat_continuous_space_with_scaling()
)
problem.metric_information.extend([
vz.MetricInformation(
name='metric1', goal=vz.ObjectiveMetricGoal.MAXIMIZE
),
vz.MetricInformation(
name='metric2', goal=vz.Objective... | problem = vz.ProblemStatement(
test_studies.flat_continuous_space_with_scaling()
)
problem.metric_information.extend([
vz.MetricInformation(
name='metric1', goal=vz.ObjectiveMetricGoal.MAXIMIZE
),
vz.MetricInformation(
name='metric2', goal=vz.Objective... | 55 | 360 | 465 | 825 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,948 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/ensemble/ensemble_design.py | 213,215 | 164,284 | play_probs | AdaptiveEnsembleDesign | 216,224 | 215,224 | @attrs.define(kw_only=True)
class AdaptiveEnsembleDesign(EnsembleDesign):
"""An EnsembleStrategy that minimizes adaptive regret."""
indices: list[int] = attrs.field()
# List of max history lengths.
max_lengths: list[int] = attrs.field()
# Base stepsize and meta_stepsize should theoretically be 1/sqrt(n) wher... | @attrs.define(kw_only=True)
class AdaptiveEnsembleDesign(EnsembleDesign):
"""An EnsembleStrategy that minimizes adaptive regret."""
indices: list[int] = attrs.field()
# List of max history lengths.
max_lengths: list[int] = attrs.field()
# Base stepsize and meta_stepsize should theoretically be 1/sqrt(n) wher... | ensemble_probs = [algo.ensemble_probs for algo in self._base_algos.values()]
weights = softmax(np.array(list(self._log_weights.values())))
final_probs = np.zeros(len(self.indices))
for weight, probs in zip(weights, ensemble_probs):
final_probs += weight * probs
return final_probs
| """Returns the probability of playing an expert."""
ensemble_probs = [algo.ensemble_probs for algo in self._base_algos.values()]
weights = softmax(np.array(list(self._log_weights.values())))
final_probs = np.zeros(len(self.indices))
for weight, probs in zip(weights, ensemble_probs):
final_prob... | 26 | 1,397 | 118 | 1,515 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,949 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/ensemble/ensemble_design.py | 88,89 | 66,102 | update | EXP3IXEnsembleDesign | 89,102 | 89,102 | @attrs.define
class EXP3IXEnsembleDesign(EnsembleDesign):
"""The EXP3-IX Algorithm that is robust against small probabilities."""
indices: list[int] = attrs.field()
stepsize: float = attrs.field(
default=1.0,
validator=[attrs.validators.instance_of(float), attrs.validators.gt(0)],
)
max_reward: f... | @attrs.define
class EXP3IXEnsembleDesign(EnsembleDesign):
"""The EXP3-IX Algorithm that is robust against small probabilities."""
indices: list[int] = attrs.field()
stepsize: float = attrs.field(
default=1.0,
validator=[attrs.validators.instance_of(float), attrs.validators.gt(0)],
)
max_reward: f... | expert_idx, reward = observation
reward = min(self.max_reward, reward)
gamma = 1.0 / np.sqrt(1.0 + len(self._history))
loss_estimator = (self.max_reward - reward) / (
self.ensemble_probs[expert_idx] + gamma
)
self._log_weights[expert_idx] += self.stepsize * (-loss_estimator)
self._h... | expert_idx, reward = observation
reward = min(self.max_reward, reward)
gamma = 1.0 / np.sqrt(1.0 + len(self._history))
loss_estimator = (self.max_reward - reward) / (
self.ensemble_probs[expert_idx] + gamma
)
self._log_weights[expert_idx] += self.stepsize * (-loss_estimator)
self._h... | 37 | 270 | 192 | 462 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,950 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/ensemble/ensemble_design.py | 132,133 | 102,164 | update | EXP3UniformEnsembleDesign | 134,156 | 133,156 | @attrs.define
class EXP3UniformEnsembleDesign(EnsembleDesign):
"""The EXP3 algorithm with uniform exploration."""
indices: list[int] = attrs.field()
stepsize: float = attrs.field(
default=1.0,
validator=[attrs.validators.instance_of(float), attrs.validators.gt(0)],
)
max_reward: float = attrs.fie... | @attrs.define
class EXP3UniformEnsembleDesign(EnsembleDesign):
"""The EXP3 algorithm with uniform exploration."""
indices: list[int] = attrs.field()
stepsize: float = attrs.field(
default=1.0,
validator=[attrs.validators.instance_of(float), attrs.validators.gt(0)],
)
max_reward: float = attrs.fie... | expert_idx, reward = observation
if not self.use_reward_estimator:
reward = min(self.max_reward, reward)
gamma = 1.0 / np.sqrt(1 + len(self._history))
if self.use_loss_formulation:
if self.use_reward_estimator:
loss_estimator = self.max_reward - reward
else:
loss_estim... | """Update history and weights."""
expert_idx, reward = observation
if not self.use_reward_estimator:
reward = min(self.max_reward, reward)
gamma = 1.0 / np.sqrt(1 + len(self._history))
if self.use_loss_formulation:
if self.use_reward_estimator:
loss_estimator = self.max_reward -... | 64 | 505 | 282 | 787 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,951 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/ensemble/ensemble_design.py | 49,51 | 45,59 | ensemble_probs | RandomEnsembleDesign | 52,54 | 51,54 | @attrs.define
class RandomEnsembleDesign(EnsembleDesign):
indices: list[int] = attrs.field()
@property
def ensemble_probs(self) -> np.ndarray:
FILL_FUNC_BODY def update(self, observation: IndexWithReward, **kwargs):
"""Updates the strategy with observation."""
pass
| @attrs.define
class RandomEnsembleDesign(EnsembleDesign):
indices: list[int] = attrs.field()
@property
def ensemble_probs(self) -> np.ndarray:
"""The probability distribution used to choose the expert."""
FILL_FUNC_BODY def update(self, observation: IndexWithReward, **kwargs):
"""Updates the strategy wi... | return np.ones(shape=len(self.indices)) / len(self.indices)
| """The probability distribution used to choose the expert."""
return np.ones(shape=len(self.indices)) / len(self.indices)
| 4 | 107 | 27 | 134 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,952 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/ensemble/ensemble_design_test.py | 34,35 | 23,108 | testEXP3IX | EnsembleDesignTest | 35,51 | 35,51 | class EnsembleDesignTest(absltest.TestCase):
def testRandom(self):
indices = [0, 1]
rewards = [(1, 1), (0, 0), (1, 1)]
strategy = ensemble_design.RandomEnsembleDesign(indices=indices)
for reward in rewards:
strategy.update(reward)
np.testing.assert_array_equal(strategy.ensemble_probs, [0... | class EnsembleDesignTest(absltest.TestCase):
def testRandom(self):
indices = [0, 1]
rewards = [(1, 1), (0, 0), (1, 1)]
strategy = ensemble_design.RandomEnsembleDesign(indices=indices)
for reward in rewards:
strategy.update(reward)
np.testing.assert_array_equal(strategy.ensemble_probs, [0... | indices = [0, 1]
rewards = [(1, 1), (0, 0), (1, 1)]
strategy = ensemble_design.EXP3IXEnsembleDesign(
indices=indices, stepsize=1.0
)
np.testing.assert_array_equal(strategy.ensemble_probs, [0.5, 0.5])
for reward in rewards:
strategy.update(reward)
probs = strategy.ensemble_pr... | indices = [0, 1]
rewards = [(1, 1), (0, 0), (1, 1)]
strategy = ensemble_design.EXP3IXEnsembleDesign(
indices=indices, stepsize=1.0
)
np.testing.assert_array_equal(strategy.ensemble_probs, [0.5, 0.5])
for reward in rewards:
strategy.update(reward)
probs = strategy.ensemble_pr... | 35 | 770 | 196 | 966 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,953 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/ensemble/ensemble_designer.py | 178,181 | 110,260 | update | EnsembleDesigner | 181,260 | 181,260 | class EnsembleDesigner(vza.Designer):
"""Ensembling of Designers."""
def __init__(
self,
designers: dict[str, vza.Designer],
ensemble_design_factory: Callable[
[list[int]], ensemble_design.EnsembleDesign
] = ensemble_design.RandomEnsembleDesign,
reward_generator: Optional[Ob... | class EnsembleDesigner(vza.Designer):
"""Ensembling of Designers."""
def __init__(
self,
designers: dict[str, vza.Designer],
ensemble_design_factory: Callable[
[list[int]], ensemble_design.EnsembleDesign
] = ensemble_design.RandomEnsembleDesign,
reward_generator: Optional[Ob... | # With no completed Trials, simply update with reward calcuations.
if not completed.trials:
for designer in self._designers.values():
designer.update(completed, all_active)
return
if self._reward_generator is None:
rewards = []
for t in completed.trials:
if t.infeasi... | # With no completed Trials, simply update with reward calcuations.
if not completed.trials:
for designer in self._designers.values():
designer.update(completed, all_active)
return
if self._reward_generator is None:
rewards = []
for t in completed.trials:
if t.infeasi... | 245 | 777 | 908 | 1,685 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,954 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/ensemble/ensemble_designer.py | 64,65 | 33,110 | curve_generator | ObjectiveRewardGenerator | 65,69 | 65,69 | @attrs.define
class ObjectiveRewardGenerator:
"""Stateful rewards generator using the objective curves of Trials."""
problem: vz.ProblemStatement = attrs.field()
all_trials: list[vz.Trial] = attrs.field(factory=list)
reward_regularization: float = attrs.field(
default=0.0,
validator=[attrs.validato... | @attrs.define
class ObjectiveRewardGenerator:
"""Stateful rewards generator using the objective curves of Trials."""
problem: vz.ProblemStatement = attrs.field()
all_trials: list[vz.Trial] = attrs.field(factory=list)
reward_regularization: float = attrs.field(
default=0.0,
validator=[attrs.validato... | return analyzers.MultiMetricCurveConverter.from_metrics_config(
self.problem.metric_information, flip_signs_for_min=True
)
| return analyzers.MultiMetricCurveConverter.from_metrics_config(
self.problem.metric_information, flip_signs_for_min=True
)
| 5 | 843 | 49 | 892 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,956 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/evolution/nsga2.py | 146,147 | 116,196 | select | NSGA2Survival | 161,196 | 147,196 | class NSGA2Survival(templates.Survival):
"""NSGA2 Survival step.
Reference: https://ieeexplore.ieee.org/document/996017
"""
def __init__(
self,
target_size: int,
*,
ranking_fn: Callable[[np.ndarray], np.ndarray] = _pareto_rank,
eviction_limit: Optional[int] = None
):
"""Ini... | class NSGA2Survival(templates.Survival):
"""NSGA2 Survival step.
Reference: https://ieeexplore.ieee.org/document/996017
"""
def __init__(
self,
target_size: int,
*,
ranking_fn: Callable[[np.ndarray], np.ndarray] = _pareto_rank,
eviction_limit: Optional[int] = None
):
"""Ini... | population = population[population.ages < self._eviction_limit]
if not population:
# return empty.
return population
selected = population.empty_like()
# Sort by the safety constraint.
if selected.cs.shape[1]:
top, border = _select_by(
_constraint_violation(population.cs... | """Applies survival process.
Sorted all points by 3-tuple
1. Descending order of safety constraint violation score. Zero
means no violations.
2. Ascending order of how many points it's dominated by.
3. Descending order of crowding distance.
Args:
population:
Returns:
Pop... | 136 | 441 | 396 | 837 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,957 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/evolution/templates.py | 164,166 | 120,201 | converter | CanonicalEvolutionDesigner | 166,168 | 166,168 | class CanonicalEvolutionDesigner(vza.PartiallySerializableDesigner,
Generic[_PopulationType, _OffspringsType]):
"""Evolution algorithm template."""
def __init__(
self,
converter: PopulationConverter[_PopulationType, _OffspringsType],
sampler: Sampler[_OffspringsTy... | class CanonicalEvolutionDesigner(vza.PartiallySerializableDesigner,
Generic[_PopulationType, _OffspringsType]):
"""Evolution algorithm template."""
def __init__(
self,
converter: PopulationConverter[_PopulationType, _OffspringsType],
sampler: Sampler[_OffspringsTy... | return self._converter
| return self._converter
| 2 | 878 | 10 | 888 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,958 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/evolution/templates.py | 56,58 | 53,62 | sample | Sampler | 59,62 | 58,62 | class Sampler(Generic[_OffspringsType], abc.ABC):
"""Creates new offsprings, typically at the start of evolutionary strategy."""
@abc.abstractmethod
def sample(self, count: int) -> _OffspringsType:
FILL_FUNC_BODY | class Sampler(Generic[_OffspringsType], abc.ABC):
"""Creates new offsprings, typically at the start of evolutionary strategy."""
@abc.abstractmethod
def sample(self, count: int) -> _OffspringsType:
"""Generate offsprings."""
FILL_FUNC_BODY |
# Temporary typevar, used for defininig Population interface.
| """Generate offsprings."""
# Temporary typevar, used for defininig Population interface.
| 8 | 89 | 21 | 110 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,959 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/optimizers/base.py | 28,29 | 1,160 | _is_positive | 29,34 | 29,34 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... | del instance, attribute
if value < 0:
raise ValueError(f'value must be positive! given: {value}')
| del instance, attribute
if value < 0:
raise ValueError(f'value must be positive! given: {value}')
| 14 | 1,537 | 39 | 1,576 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... | |
46,960 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/optimizers/base.py | 137,146 | 115,160 | optimize | BranchThenOptimizer | 146,160 | 146,160 | @attr.frozen
class BranchThenOptimizer(GradientFreeOptimizer):
"""Optimizes a function by first choosing a branch and then apply Optimizer.
Attributes:
_branch_selector: Selects all conditional parent values
_optimizer_factory: Creates an optimizer in the flat (non-conditional) after
branch selector ... | @attr.frozen
class BranchThenOptimizer(GradientFreeOptimizer):
"""Optimizes a function by first choosing a branch and then apply Optimizer.
Attributes:
_branch_selector: Selects all conditional parent values
_optimizer_factory: Creates an optimizer in the flat (non-conditional) after
branch selector ... | # If there are conditional branches, use Vizier's default branch
# selection mechanism.
branches = self._branch_selector.select_branches(count)
suggestions = []
optimizer = self._optimizer_factory()
for branch in branches:
subproblem = attr.evolve(problem, search_space=branch.search_space)... | # If there are conditional branches, use Vizier's default branch
# selection mechanism.
branches = self._branch_selector.select_branches(count)
suggestions = []
optimizer = self._optimizer_factory()
for branch in branches:
subproblem = attr.evolve(problem, search_space=branch.search_space)... | 42 | 374 | 168 | 542 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,961 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/optimizers/designer_optimizer_test.py | 41,42 | 28,54 | test_bi_objective | DesignerOptimizerTest | 42,54 | 42,54 | class DesignerOptimizerTest(absltest.TestCase):
def test_smoke(self):
problem = vz.ProblemStatement()
problem.search_space.root.add_float_param('a', 0, 1)
designer_factory = quasi_random.QuasiRandomDesigner.from_problem
optimizer_test_utils.assert_passes_on_random_single_metric_function(
sel... | class DesignerOptimizerTest(absltest.TestCase):
def test_smoke(self):
problem = vz.ProblemStatement()
problem.search_space.root.add_float_param('a', 0, 1)
designer_factory = quasi_random.QuasiRandomDesigner.from_problem
optimizer_test_utils.assert_passes_on_random_single_metric_function(
sel... | problem = vz.ProblemStatement()
problem.search_space.root.add_float_param('a', 0, 1)
designer_factory = nsga2.NSGA2Designer
optimizer_test_utils.assert_passes_on_random_multi_metric_function(
self,
problem.search_space,
designer_optimizer.DesignerAsOptimizer(designer_factory),
... | problem = vz.ProblemStatement()
problem.search_space.root.add_float_param('a', 0, 1)
designer_factory = nsga2.NSGA2Designer
optimizer_test_utils.assert_passes_on_random_multi_metric_function(
self,
problem.search_space,
designer_optimizer.DesignerAsOptimizer(designer_factory),
... | 15 | 169 | 133 | 302 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,962 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/optimizers/random_vectorized_optimizer_test.py | 30,31 | 27,52 | test_random_optimizer | RandomVectorizedOptimizerTest | 31,44 | 31,44 | class RandomVectorizedOptimizerTest(absltest.TestCase):
"""Tests for random_vectorized_optimizer."""
def test_random_optimizer(self):
FILL_FUNC_BODY def test_random_optimizer_factory(self):
random_optimizer_factory = rvo.create_random_optimizer_factory(
max_evaluations=100, suggestion_batch_size=10
... | class RandomVectorizedOptimizerTest(absltest.TestCase):
"""Tests for random_vectorized_optimizer."""
def test_random_optimizer(self):
FILL_FUNC_BODY def test_random_optimizer_factory(self):
random_optimizer_factory = rvo.create_random_optimizer_factory(
max_evaluations=100, suggestion_batch_size=10
... | problem = vz.ProblemStatement()
problem.search_space.root.add_float_param('f1', 0.0, 10.0)
problem.search_space.root.add_float_param('f2', 0.0, 10.0)
converter = converters.TrialToModelInputConverter.from_problem(problem)
score_fn = lambda x, _: np.sum(x.continuous.padded_array, axis=(-1, -2))
n... | problem = vz.ProblemStatement()
problem.search_space.root.add_float_param('f1', 0.0, 10.0)
problem.search_space.root.add_float_param('f2', 0.0, 10.0)
converter = converters.TrialToModelInputConverter.from_problem(problem)
score_fn = lambda x, _: np.sum(x.continuous.padded_array, axis=(-1, -2))
n... | 41 | 158 | 261 | 419 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,963 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/policies/designer_policy.py | 77,78 | 40,126 | suggest | DesignerPolicy | 78,114 | 78,114 | class DesignerPolicy(pythia.Policy):
"""Wraps a Designer into a pythia Policy.
> IMPORTANT: If your Designer class is (partially) serializable, use
> (Partially)SerializableDesignerPolicy instead.
When a Designer cannot be (partially) serialized, we create a new Designer
instance at the start of each `sugge... | class DesignerPolicy(pythia.Policy):
"""Wraps a Designer into a pythia Policy.
> IMPORTANT: If your Designer class is (partially) serializable, use
> (Partially)SerializableDesignerPolicy instead.
When a Designer cannot be (partially) serialized, we create a new Designer
instance at the start of each `sugge... | logging.info(
'Processing request for study guid %s, with max trial id %s',
request.study_guid,
request.max_trial_id,
)
designer = self._designer_factory(request.study_config)
logging.info(
'Study guid %s: Designer created. Getting trials from Vizier service.',
re... | logging.info(
'Processing request for study guid %s, with max trial id %s',
request.study_guid,
request.max_trial_id,
)
designer = self._designer_factory(request.study_config)
logging.info(
'Study guid %s: Designer created. Getting trials from Vizier service.',
re... | 99 | 471 | 439 | 910 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,964 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/policies/designer_policy.py | 351,352 | 347,364 | _restore_designer | InRamDesignerPolicy | 352,360 | 352,360 | class InRamDesignerPolicy(_SerializableDesignerPolicyBase[vza.Designer]):
"""Wraps a Designer in RAM for efficient updates()."""
# Override restore() to simply return the designer.
def _restore_designer(self, designer_metadata: vz.Metadata) -> vza.Designer:
FILL_FUNC_BODY def dump(self) -> vz.Metadata:
retu... | class InRamDesignerPolicy(_SerializableDesignerPolicyBase[vza.Designer]):
"""Wraps a Designer in RAM for efficient updates()."""
# Override restore() to simply return the designer.
def _restore_designer(self, designer_metadata: vz.Metadata) -> vza.Designer:
FILL_FUNC_BODY def dump(self) -> vz.Metadata:
retu... | if self._designer is None:
raise ValueError(
'`self._designer` has not been initialized!'
' Call self._initialize_designer(..) first.'
)
return self._designer
# Override dump() since this is in ram.
| if self._designer is None:
raise ValueError(
'`self._designer` has not been initialized!'
' Call self._initialize_designer(..) first.'
)
return self._designer
# Override dump() since this is in ram.
| 26 | 115 | 78 | 193 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,965 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/policies/designer_policy.py | 369,372 | 364,377 | _restore_designer | PartiallySerializableDesignerPolicy | 372,377 | 372,377 | class PartiallySerializableDesignerPolicy(
_SerializableDesignerPolicyBase[vza.PartiallySerializableDesigner]
):
"""Wraps a PartiallySerializableDesigner."""
def _restore_designer(
self, designer_metadata: vz.Metadata
) -> vza.PartiallySerializableDesigner:
FILL_FUNC_BODY | class PartiallySerializableDesignerPolicy(
_SerializableDesignerPolicyBase[vza.PartiallySerializableDesigner]
):
"""Wraps a PartiallySerializableDesigner."""
def _restore_designer(
self, designer_metadata: vz.Metadata
) -> vza.PartiallySerializableDesigner:
FILL_FUNC_BODY | designer = self._designer_factory(self._problem_statement)
designer.load(designer_metadata)
return designer
| designer = self._designer_factory(self._problem_statement)
designer.load(designer_metadata)
return designer
| 6 | 93 | 39 | 132 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,966 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/policies/designer_policy.py | 412,415 | 377,416 | _restore_designer | SerializableDesignerPolicy | 415,416 | 415,416 | class SerializableDesignerPolicy(
_SerializableDesignerPolicyBase[vza.SerializableDesigner]
):
"""Wraps a SerializableDesigner."""
def __init__(
self,
problem_statement: vz.ProblemStatement,
supporter: pythia.PolicySupporter,
designer_factory: vza.DesignerFactory[vza.SerializableDesigne... | class SerializableDesignerPolicy(
_SerializableDesignerPolicyBase[vza.SerializableDesigner]
):
"""Wraps a SerializableDesigner."""
def __init__(
self,
problem_statement: vz.ProblemStatement,
supporter: pythia.PolicySupporter,
designer_factory: vza.DesignerFactory[vza.SerializableDesigne... | return self._designer_cls.recover(designer_metadata)
| return self._designer_cls.recover(designer_metadata)
| 2 | 342 | 19 | 361 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,968 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/policies/random_policy.py | 42,45 | 29,70 | early_stop | RandomPolicy | 46,70 | 45,70 | class RandomPolicy(pythia.Policy):
"""A policy that picks random hyper-parameter values."""
def __init__(self, policy_supporter: pythia.PolicySupporter):
self._policy_supporter = policy_supporter
def suggest(self, request: pythia.SuggestRequest) -> pythia.SuggestDecision:
"""Gets number of Trials to pro... | class RandomPolicy(pythia.Policy):
"""A policy that picks random hyper-parameter values."""
def __init__(self, policy_supporter: pythia.PolicySupporter):
self._policy_supporter = policy_supporter
def suggest(self, request: pythia.SuggestRequest) -> pythia.SuggestDecision:
"""Gets number of Trials to pro... | decisions = []
all_active_trials = self._policy_supporter.GetTrials(
study_guid=request.study_guid,
status_matches=pyvizier.TrialStatus.ACTIVE,
)
trial_to_stop_id = None
if all_active_trials:
trial_to_stop_id = random.choice(all_active_trials).id
decisions.append(
... | """Selects a random ACTIVE/PENDING trial to stop from datastore."""
decisions = []
all_active_trials = self._policy_supporter.GetTrials(
study_guid=request.study_guid,
status_matches=pyvizier.TrialStatus.ACTIVE,
)
trial_to_stop_id = None
if all_active_trials:
trial_to_stop... | 46 | 225 | 280 | 505 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,969 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/policies/trial_caches.py | 80,81 | 32,136 | get_newly_completed_trials | IdDeduplicatingTrialLoader | 82,103 | 81,103 | @attrs.define
class IdDeduplicatingTrialLoader(serializable.PartiallySerializable):
"""Wrapper around PolicySupporter to avoid reloading completed trials.
get_newly_completed_trials() only return the trials completed since the
last call.
get_active_trials() is just a convienence wrapper.
"""
# TODO: Stor... | @attrs.define
class IdDeduplicatingTrialLoader(serializable.PartiallySerializable):
"""Wrapper around PolicySupporter to avoid reloading completed trials.
get_newly_completed_trials() only return the trials completed since the
last call.
get_active_trials() is just a convienence wrapper.
"""
# TODO: Stor... | if len(self._incorporated_completed_trial_ids) == max_trial_id:
# no trials need to be loaded.
return []
all_trial_ids = set(range(1, max_trial_id + 1))
# Exclude completed trials that were already passed to the designer.
trial_ids_to_load = all_trial_ids - self._incorporated_completed_trial... | """Returns trials completed between the last call and max_trial_id."""
if len(self._incorporated_completed_trial_ids) == max_trial_id:
# no trials need to be loaded.
return []
all_trial_ids = set(range(1, max_trial_id + 1))
# Exclude completed trials that were already passed to the designer.... | 78 | 965 | 306 | 1,271 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,970 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/random/random_sample.py | 79,80 | 1,127 | get_closest_element | 81,86 | 80,86 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... | gaps = [abs(x - value) for x in array]
closest_idx = min(enumerate(gaps), key=lambda x: x[1])[0]
return array[closest_idx]
| """Finds closest element in array to value."""
gaps = [abs(x - value) for x in array]
closest_idx = min(enumerate(gaps), key=lambda x: x[1])[0]
return array[closest_idx]
| 17 | 1,302 | 58 | 1,360 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... | |
46,971 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/algorithms/random/random_sample_test.py | 104,105 | 27,128 | test_sample_value_categorical | RandomSampleTest | 105,110 | 105,110 | class RandomSampleTest(parameterized.TestCase):
def setUp(self):
super(RandomSampleTest, self).setUp()
self.rng = np.random.default_rng(0)
@parameterized.named_parameters(
dict(testcase_name='prob=0', value=0.2, target=0.0),
dict(testcase_name='prob=1.0', value=-0.5, target=0.0),
dict(te... | class RandomSampleTest(parameterized.TestCase):
def setUp(self):
super(RandomSampleTest, self).setUp()
self.rng = np.random.default_rng(0)
@parameterized.named_parameters(
dict(testcase_name='prob=0', value=0.2, target=0.0),
dict(testcase_name='prob=1.0', value=-0.5, target=0.0),
dict(te... | categorical_param = vz.ParameterConfig.factory(
'categorical', feasible_values=['a', 'b', 'c'])
value = random_sample._sample_value(self.rng, categorical_param)
self.assertIn(value, ['a', 'b', 'c'])
| categorical_param = vz.ParameterConfig.factory(
'categorical', feasible_values=['a', 'b', 'c'])
value = random_sample._sample_value(self.rng, categorical_param)
self.assertIn(value, ['a', 'b', 'c'])
| 15 | 1,630 | 81 | 1,711 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,973 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/analyzers/convergence_curve.py | 636,641 | 558,689 | standardize_curves | ConvergenceComparator | 656,689 | 641,689 | @attr.define
class ConvergenceComparator(abc.ABC):
"""Base class for convergence curve compartors.
Attributes:
baseline_curve: The baseline ConvergenceCurve.
compared_curve: The compared ConvergenceCurve.
baseline_quantile: Quantile in [0, 1] of the batched baseline curve to use
for efficiency co... | @attr.define
class ConvergenceComparator(abc.ABC):
"""Base class for convergence curve compartors.
Attributes:
baseline_curve: The baseline ConvergenceCurve.
compared_curve: The compared ConvergenceCurve.
baseline_quantile: Quantile in [0, 1] of the batched baseline curve to use
for efficiency co... | # Align the curves while keeping each ys.
align_baseline_curve, align_compared_curve = ConvergenceCurve.align_xs(
[self._baseline_curve, self._compared_curve],
interpolate_repeats=False,
keep_curves_separate=True,
)
# Adjust sign to increasing.
baseline_ys = self._sign * alig... | """Standardize convergence curves.
1. Align xs and keep each ys.
2. Convert curves to INCREASING.
3. Apply quantiles and impute NaN (optional).
4. Remove values where xs < xs_cutoff.
Args:
xs_cutoff: The xs value before which values are ignored.
apply_quantiles: Whether to compute ... | 132 | 1,050 | 546 | 1,596 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,975 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/analyzers/convergence_curve.py | 282,283 | 254,342 | convert | ConvergenceCurveConverter | 284,334 | 283,334 | @attr.define
class ConvergenceCurveConverter(StatefulCurveConverter):
"""Converter for Trial sequence to ConvergenceCurve.
Attributes:
metric_information: Information of relevant metric.
flip_signs_for_min: If True, flips the signs of metric values to always
maximize. Useful when desiring all i... | @attr.define
class ConvergenceCurveConverter(StatefulCurveConverter):
"""Converter for Trial sequence to ConvergenceCurve.
Attributes:
metric_information: Information of relevant metric.
flip_signs_for_min: If True, flips the signs of metric values to always
maximize. Useful when desiring all i... | if not trials:
raise ValueError(f'No trials provided {trials}')
yvals = [self._best_yval]
xvals = [self._cumulative_cost]
candidates = []
for i in range(0, len(trials), self.batch_size):
batch = trials[i : i + self.batch_size]
for trial in batch:
if self.measurements_type... | """Returns ConvergenceCurve with a single curve."""
if not trials:
raise ValueError(f'No trials provided {trials}')
yvals = [self._best_yval]
xvals = [self._cumulative_cost]
candidates = []
for i in range(0, len(trials), self.batch_size):
batch = trials[i : i + self.batch_size]
... | 129 | 526 | 628 | 1,154 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,976 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/analyzers/convergence_curve.py | 391,392 | 342,462 | convert | HypervolumeCurveConverter | 393,462 | 392,462 | class HypervolumeCurveConverter(StatefulCurveConverter):
"""Converts Trials to cumulative hypervolume curve for multiobjective."""
def __init__(
self,
metric_informations: Sequence[pyvizier.MetricInformation],
*,
reference_value: Optional[np.ndarray] = None,
num_vectors: int = 10000,
... | class HypervolumeCurveConverter(StatefulCurveConverter):
"""Converts Trials to cumulative hypervolume curve for multiobjective."""
def __init__(
self,
metric_informations: Sequence[pyvizier.MetricInformation],
*,
reference_value: Optional[np.ndarray] = None,
num_vectors: int = 10000,
... | # Returns a len(trials) x number of metrics np.ndarray.
if not trials:
raise ValueError(f'No trials provided {trials}')
metrics = self._converter.to_labels_array(trials)
if self._origin_value is None:
# Set origin to the minimum of finite values.
origin = np.zeros(shape=(metrics.shape... | """Returns ConvergenceCurve with a curve of shape 1 x len(trials)."""
# Returns a len(trials) x number of metrics np.ndarray.
if not trials:
raise ValueError(f'No trials provided {trials}')
metrics = self._converter.to_labels_array(trials)
if self._origin_value is None:
# Set origin to ... | 199 | 570 | 928 | 1,498 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,977 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/analyzers/convergence_curve.py | 772,773 | 703,826 | score | LogEfficiencyConvergenceCurveComparator | 783,826 | 773,826 | @attr.define
class LogEfficiencyConvergenceCurveComparator(ConvergenceComparator):
"""Comparator methods for ConvergenceCurves.
Methods in this class generally return comparison metrics for a compared curve
against a baseline curve. Only works for curves with INCREASING or DECREASING
trend.
Example usage:
... | @attr.define
class LogEfficiencyConvergenceCurveComparator(ConvergenceComparator):
"""Comparator methods for ConvergenceCurves.
Methods in this class generally return comparison metrics for a compared curve
against a baseline curve. Only works for curves with INCREASING or DECREASING
trend.
Example usage:
... | baseline_curve = ConvergenceCurve.align_xs(
[self._baseline_curve], interpolate_repeats=True
)[0]
compared_curve = ConvergenceCurve.align_xs(
[self._compared_curve], interpolate_repeats=True
)[0]
# Combined curve (as the baseline) becomes the y-values at which
# Trial efficiency ... | """Gets a finalized log efficiency score.
The compared curve should approximately use exp(-score)% Trials compared to
the baseline curve. Note that a high positive score demonstrates that the
compared curve uses less Trials and is better than the baseline.
Returns:
Sample efficiency score. T... | 116 | 846 | 612 | 1,458 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,978 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/analyzers/convergence_curve.py | 502,503 | 462,514 | convert | MultiMetricCurveConverter | 504,514 | 503,514 | @attr.define(init=True)
class MultiMetricCurveConverter(StatefulCurveConverter):
"""Converts Trials to cumulative convergence curve for all multimetric studies.
Attributes:
metrics_config: Metrics config of the whole study.
converter: Wrapped curve converter.
"""
metrics_config: pyvizier.MetricsCo... | @attr.define(init=True)
class MultiMetricCurveConverter(StatefulCurveConverter):
"""Converts Trials to cumulative convergence curve for all multimetric studies.
Attributes:
metrics_config: Metrics config of the whole study.
converter: Wrapped curve converter.
"""
metrics_config: pyvizier.MetricsCo... | if not trials:
raise ValueError(f'No trials provided {trials}')
# Add safety understanding by setting unsafe Trials to inf/-inf.
safety_checker = multimetric.SafetyChecker(self.metrics_config)
warped_trials = safety_checker.warp_unsafe_trials(copy.deepcopy(trials))
return self.converter.conv... | """Returns ConvergenceCurve with a curve of shape 1 x len(trials)."""
if not trials:
raise ValueError(f'No trials provided {trials}')
# Add safety understanding by setting unsafe Trials to inf/-inf.
safety_checker = multimetric.SafetyChecker(self.metrics_config)
warped_trials = safety_checker... | 26 | 449 | 123 | 572 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,979 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/analyzers/convergence_curve.py | 979,980 | 968,990 | score | OptimalityGapGainComparator | 981,985 | 980,985 | class OptimalityGapGainComparator(ConvergenceComparator):
"""Comparator method based on optimality gap gain.
The optimality gap gain ('compared' - 'baseline') is normalized by the
'baseline' absolute optimality gap and then truncated.
"""
min_value: float = -0.5
max_value: float = 1.0
eps: float = 0.000... | class OptimalityGapGainComparator(ConvergenceComparator):
"""Comparator method based on optimality gap gain.
The optimality gap gain ('compared' - 'baseline') is normalized by the
'baseline' absolute optimality gap and then truncated.
"""
min_value: float = -0.5
max_value: float = 1.0
eps: float = 0.000... | baseline_ys, compared_ys = self.standardize_curves()
d = (compared_ys[-1] - baseline_ys[-1]) / (abs(baseline_ys[-1]) + self.eps)
return min(max(d, self.min_value), self.max_value)
| """Computes the normalized simple regert score."""
baseline_ys, compared_ys = self.standardize_curves()
d = (compared_ys[-1] - baseline_ys[-1]) / (abs(baseline_ys[-1]) + self.eps)
return min(max(d, self.min_value), self.max_value)
| 17 | 184 | 79 | 263 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,980 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/analyzers/convergence_curve.py | 963,964 | 954,968 | curve | OptimalityGapWinRateComparator | 965,968 | 964,968 | @attr.define
class OptimalityGapWinRateComparator(ConvergenceComparator):
"""Comparator method based on win-rate of the optimality gap."""
def score(self):
"""Computes the normalized simple regert score."""
baseline_ys, compared_ys = self.standardize_curves()
return float(compared_ys[-1] > baseline_ys[... | @attr.define
class OptimalityGapWinRateComparator(ConvergenceComparator):
"""Comparator method based on win-rate of the optimality gap."""
def score(self):
"""Computes the normalized simple regert score."""
baseline_ys, compared_ys = self.standardize_curves()
return float(compared_ys[-1] > baseline_ys[... | raise NotImplementedError('Curve not yet implemented.')
| """Returns a score curve for each xs."""
raise NotImplementedError('Curve not yet implemented.')
| 5 | 137 | 21 | 158 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,981 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/analyzers/convergence_curve.py | 878,879 | 826,907 | score | PercentageBetterConvergenceCurveComparator | 894,903 | 879,903 | @attr.define
class PercentageBetterConvergenceCurveComparator(ConvergenceComparator):
"""Comparator method based on percentage better.
PercentageBetter is the average percentage of steps that one curve is better
than the other.
For example, assuming a study with 100 trials, a score of 0.07 means that on
ave... | @attr.define
class PercentageBetterConvergenceCurveComparator(ConvergenceComparator):
"""Comparator method based on percentage better.
PercentageBetter is the average percentage of steps that one curve is better
than the other.
For example, assuming a study with 100 trials, a score of 0.07 means that on
ave... | baseline_ys, compared_ys = self.standardize_curves(self._xs_cutoff)
baseline_compared_score = self._compute_directional_score(
baseline_ys, compared_ys
)
compared_baseline_score = self._compute_directional_score(
compared_ys, baseline_ys
)
return baseline_compared_score - compare... | """Computes the percentage better score.
The score has two components:
(a) sub-score quantifying how much 'compared' domniates 'baseline'.
(b) sub-score quantifying how much 'baseline' domniates 'compared'.
The final score is (a) - (b).
Returns:
The normalized percentage better converge... | 20 | 765 | 111 | 876 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,982 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/analyzers/convergence_curve.py | 534,535 | 514,558 | convert | RestartingCurveConverter | 535,558 | 535,558 | @attr.define(init=True)
class RestartingCurveConverter(StatefulCurveConverter):
"""StatefulConverter that restarts the underlying stateful converters."""
converter_factory: Callable[[], StatefulCurveConverter] = attr.field()
# The minimum number of Trials needed before restarts occur.
restart_min_trials: int =... | @attr.define(init=True)
class RestartingCurveConverter(StatefulCurveConverter):
"""StatefulConverter that restarts the underlying stateful converters."""
converter_factory: Callable[[], StatefulCurveConverter] = attr.field()
# The minimum number of Trials needed before restarts occur.
restart_min_trials: int =... | if self._converter is None:
self._converter = self.converter_factory()
if self._all_trials:
self._converter.convert(self._all_trials)
curve = self._converter.convert(trials)
self._all_trials.extend(trials)
if len(self._all_trials) < self.restart_min_trials:
return curve
#... | if self._converter is None:
self._converter = self.converter_factory()
if self._all_trials:
self._converter.convert(self._all_trials)
curve = self._converter.convert(trials)
self._all_trials.extend(trials)
if len(self._all_trials) < self.restart_min_trials:
return curve
#... | 57 | 284 | 270 | 554 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,983 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/analyzers/convergence_curve.py | 928,929 | 907,954 | curve | WinRateConvergenceCurveComparator | 930,954 | 929,954 | @attr.define
class WinRateConvergenceCurveComparator(ConvergenceComparator):
"""Comparator method based on convergence curves simple win rate comparison.
The comparator has two modes of comparing convergence curves:
1. Pairwise - Compare all pairs of repeated convergence curves and then
compute the mean win... | @attr.define
class WinRateConvergenceCurveComparator(ConvergenceComparator):
"""Comparator method based on convergence curves simple win rate comparison.
The comparator has two modes of comparing convergence curves:
1. Pairwise - Compare all pairs of repeated convergence curves and then
compute the mean win... | if self.comparison_mode == 'pairwise':
baseline_ys, compared_ys = self.standardize_curves(apply_quantiles=False)
# Compares all pairs of compared to baseline curve.
all_comparisons = np.apply_along_axis(
lambda base: np.mean(compared_ys > base, axis=0)
+ 0.5 * np.mean(base == c... | """Computes the curve that represents the average win rate."""
if self.comparison_mode == 'pairwise':
baseline_ys, compared_ys = self.standardize_curves(apply_quantiles=False)
# Compares all pairs of compared to baseline curve.
all_comparisons = np.apply_along_axis(
lambda base: np.m... | 100 | 271 | 390 | 661 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,985 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/analyzers/convergence_curve_test.py | 176,209 | 148,264 | test_convert_with_batch_size | ConvergenceCurveConverterTest | 209,219 | 209,219 | class ConvergenceCurveConverterTest(parameterized.TestCase):
@parameterized.named_parameters(
('maximize', pyvizier.ObjectiveMetricGoal.MAXIMIZE, [[2, 2, 3]]),
('minimize', pyvizier.ObjectiveMetricGoal.MINIMIZE, [[2, 1, 1]]),
)
def test_convert_basic(self, goal, expected):
trials = _gen_trials([2... | class ConvergenceCurveConverterTest(parameterized.TestCase):
@parameterized.named_parameters(
('maximize', pyvizier.ObjectiveMetricGoal.MAXIMIZE, [[2, 2, 3]]),
('minimize', pyvizier.ObjectiveMetricGoal.MINIMIZE, [[2, 1, 1]]),
)
def test_convert_basic(self, goal, expected):
trials = _gen_trials([2... | trials = _gen_trials(values)
generator = convergence.ConvergenceCurveConverter(
pyvizier.MetricInformation(name='', goal=goal),
flip_signs_for_min=False,
batch_size=batch_size,
)
curve = generator.convert(trials)
np.testing.assert_array_equal(curve.xs, list(range(1, len(trial... | trials = _gen_trials(values)
generator = convergence.ConvergenceCurveConverter(
pyvizier.MetricInformation(name='', goal=goal),
flip_signs_for_min=False,
batch_size=batch_size,
)
curve = generator.convert(trials)
np.testing.assert_array_equal(curve.xs, list(range(1, len(trial... | 21 | 1,515 | 146 | 1,661 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,986 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/analyzers/convergence_curve_test.py | 40,41 | 38,148 | test_align_xs_merge_ys_on_different_lengths | ConvergenceCurveTest | 41,58 | 41,58 | class ConvergenceCurveTest(absltest.TestCase):
def test_align_xs_merge_ys_on_different_lengths(self):
FILL_FUNC_BODY def test_align_xs_merge_ys_on_distinct_xvalues(self):
c1 = convergence.ConvergenceCurve(
xs=np.array([1, 3, 4]),
ys=np.array([[2, 1, 1]]),
trend=convergence.ConvergenceCur... | class ConvergenceCurveTest(absltest.TestCase):
def test_align_xs_merge_ys_on_different_lengths(self):
FILL_FUNC_BODY def test_align_xs_merge_ys_on_distinct_xvalues(self):
c1 = convergence.ConvergenceCurve(
xs=np.array([1, 3, 4]),
ys=np.array([[2, 1, 1]]),
trend=convergence.ConvergenceCur... | c1 = convergence.ConvergenceCurve(
xs=np.array([1, 2, 3]),
ys=np.array([[2, 1, 1]]),
trend=convergence.ConvergenceCurve.YTrend.DECREASING,
)
c2 = convergence.ConvergenceCurve(
xs=np.array([1]),
ys=np.array([[3]]),
trend=convergence.ConvergenceCurve.YTrend.DECR... | c1 = convergence.ConvergenceCurve(
xs=np.array([1, 2, 3]),
ys=np.array([[2, 1, 1]]),
trend=convergence.ConvergenceCurve.YTrend.DECREASING,
)
c2 = convergence.ConvergenceCurve(
xs=np.array([1]),
ys=np.array([[3]]),
trend=convergence.ConvergenceCurve.YTrend.DECR... | 35 | 1,429 | 255 | 1,684 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,988 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/analyzers/convergence_curve_test.py | 839,840 | 746,887 | test_effiency_score_symmetry | LogEfficiencyConvergenceComparatorTest | 840,848 | 840,848 | class LogEfficiencyConvergenceComparatorTest(absltest.TestCase):
def setUp(self):
super(LogEfficiencyConvergenceComparatorTest, self).setUp()
xs = np.array(range(0, 20))
xs_t = xs.reshape(1, len(xs))
self._baseline_curve = convergence.ConvergenceCurve(
xs=xs,
ys=np.exp(np.array([-0.9,... | class LogEfficiencyConvergenceComparatorTest(absltest.TestCase):
def setUp(self):
super(LogEfficiencyConvergenceComparatorTest, self).setUp()
xs = np.array(range(0, 20))
xs_t = xs.reshape(1, len(xs))
self._baseline_curve = convergence.ConvergenceCurve(
xs=xs,
ys=np.exp(np.array([-0.9,... | base_score = convergence.LogEfficiencyConvergenceCurveComparator(
baseline_curve=self._baseline_curve, compared_curve=self._better_curves
).score()
reversed_score = convergence.LogEfficiencyConvergenceCurveComparator(
baseline_curve=self._better_curves, compared_curve=self._baseline_curve
... | base_score = convergence.LogEfficiencyConvergenceCurveComparator(
baseline_curve=self._baseline_curve, compared_curve=self._better_curves
).score()
reversed_score = convergence.LogEfficiencyConvergenceCurveComparator(
baseline_curve=self._better_curves, compared_curve=self._baseline_curve
... | 15 | 1,850 | 141 | 1,991 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,989 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/analyzers/convergence_curve_test.py | 517,518 | 480,560 | test_convert_multiobjective | MultiMetricCurveConverterTest | 518,560 | 518,560 | class MultiMetricCurveConverterTest(parameterized.TestCase):
def test_convert_single_objective(self):
generator = convergence.MultiMetricCurveConverter.from_metrics_config(
metrics_config=pyvizier.MetricsConfig([
pyvizier.MetricInformation(
name='min', goal=pyvizier.ObjectiveM... | class MultiMetricCurveConverterTest(parameterized.TestCase):
def test_convert_single_objective(self):
generator = convergence.MultiMetricCurveConverter.from_metrics_config(
metrics_config=pyvizier.MetricsConfig([
pyvizier.MetricInformation(
name='min', goal=pyvizier.ObjectiveM... | generator = convergence.MultiMetricCurveConverter.from_metrics_config(
metrics_config=pyvizier.MetricsConfig([
pyvizier.MetricInformation(
name='max', goal=pyvizier.ObjectiveMetricGoal.MAXIMIZE
),
pyvizier.MetricInformation(
name='min', goa... | generator = convergence.MultiMetricCurveConverter.from_metrics_config(
metrics_config=pyvizier.MetricsConfig([
pyvizier.MetricInformation(
name='max', goal=pyvizier.ObjectiveMetricGoal.MAXIMIZE
),
pyvizier.MetricInformation(
name='min', goa... | 69 | 447 | 515 | 962 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,990 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/analyzers/convergence_curve_test.py | 960,987 | 958,995 | test_score | OptimalityGapGainComparatorTest | 987,995 | 987,995 | class OptimalityGapGainComparatorTest(parameterized.TestCase):
@parameterized.parameters(
{
'ys1': np.array([[1, 4, 8, 10, 12]]),
'ys2': np.array([[2, 5, 6, 8, 10]]),
'trend': convergence.ConvergenceCurve.YTrend.INCREASING,
'res': (10 - 12) / 12,
},
{
... | class OptimalityGapGainComparatorTest(parameterized.TestCase):
@parameterized.parameters(
{
'ys1': np.array([[1, 4, 8, 10, 12]]),
'ys2': np.array([[2, 5, 6, 8, 10]]),
'trend': convergence.ConvergenceCurve.YTrend.INCREASING,
'res': (10 - 12) / 12,
},
{
... | xs1 = np.arange(ys1.shape[1])
xs2 = np.arange(ys2.shape[1])
curve1 = convergence.ConvergenceCurve(xs=xs1, ys=ys1, trend=trend)
curve2 = convergence.ConvergenceCurve(xs=xs2, ys=ys2, trend=trend)
comparator = convergence.OptimalityGapGainComparator(curve1, curve2)
self.assertAlmostEqual(comparator... | xs1 = np.arange(ys1.shape[1])
xs2 = np.arange(ys2.shape[1])
curve1 = convergence.ConvergenceCurve(xs=xs1, ys=ys1, trend=trend)
curve2 = convergence.ConvergenceCurve(xs=xs2, ys=ys2, trend=trend)
comparator = convergence.OptimalityGapGainComparator(curve1, curve2)
self.assertAlmostEqual(comparator... | 23 | 452 | 155 | 607 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,991 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/analyzers/convergence_curve_test.py | 997,1024 | 995,1032 | test_score | OptimalityGapWinRateComparatorTest | 1024,1032 | 1024,1032 | class OptimalityGapWinRateComparatorTest(parameterized.TestCase):
@parameterized.parameters(
{
'ys1': np.array([[1, 4, 8, 10, 12]]),
'ys2': np.array([[2, 5, 6, 8, 10]]),
'trend': convergence.ConvergenceCurve.YTrend.INCREASING,
'res': 0.0,
},
{
'ys1'... | class OptimalityGapWinRateComparatorTest(parameterized.TestCase):
@parameterized.parameters(
{
'ys1': np.array([[1, 4, 8, 10, 12]]),
'ys2': np.array([[2, 5, 6, 8, 10]]),
'trend': convergence.ConvergenceCurve.YTrend.INCREASING,
'res': 0.0,
},
{
'ys1'... | xs1 = np.arange(ys1.shape[1])
xs2 = np.arange(ys2.shape[1])
curve1 = convergence.ConvergenceCurve(xs=xs1, ys=ys1, trend=trend)
curve2 = convergence.ConvergenceCurve(xs=xs2, ys=ys2, trend=trend)
comparator = convergence.OptimalityGapWinRateComparator(curve1, curve2)
self.assertEqual(comparator.sc... | xs1 = np.arange(ys1.shape[1])
xs2 = np.arange(ys2.shape[1])
curve1 = convergence.ConvergenceCurve(xs=xs1, ys=ys1, trend=trend)
curve2 = convergence.ConvergenceCurve(xs=xs2, ys=ys2, trend=trend)
comparator = convergence.OptimalityGapWinRateComparator(curve1, curve2)
self.assertEqual(comparator.sc... | 22 | 437 | 143 | 580 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,992 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/analyzers/convergence_curve_test.py | 926,944 | 924,958 | test_score | PercentageBetterConvergenceComparatorTest | 944,958 | 944,958 | class PercentageBetterConvergenceComparatorTest(parameterized.TestCase):
@parameterized.parameters(
{
'ys1': np.array([[1, 2, 3, 4], [1, 2, 3, 4]]),
'ys2': np.array([[-1, 20, 30, 70]]),
'res': 0.5,
},
{
'ys1': np.array([[1, 2, 3, 4]]),
'ys2': np.arr... | class PercentageBetterConvergenceComparatorTest(parameterized.TestCase):
@parameterized.parameters(
{
'ys1': np.array([[1, 2, 3, 4], [1, 2, 3, 4]]),
'ys2': np.array([[-1, 20, 30, 70]]),
'res': 0.5,
},
{
'ys1': np.array([[1, 2, 3, 4]]),
'ys2': np.arr... | xs1 = np.arange(ys1.shape[1])
xs2 = np.arange(ys2.shape[1])
curve1 = convergence.ConvergenceCurve(
xs=xs1, ys=ys1, trend=convergence.ConvergenceCurve.YTrend.INCREASING
)
curve2 = convergence.ConvergenceCurve(
xs=xs2, ys=ys2, trend=convergence.ConvergenceCurve.YTrend.INCREASING
)
... | xs1 = np.arange(ys1.shape[1])
xs2 = np.arange(ys2.shape[1])
curve1 = convergence.ConvergenceCurve(
xs=xs1, ys=ys1, trend=convergence.ConvergenceCurve.YTrend.INCREASING
)
curve2 = convergence.ConvergenceCurve(
xs=xs2, ys=ys2, trend=convergence.ConvergenceCurve.YTrend.INCREASING
)
... | 28 | 270 | 197 | 467 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,993 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/analyzers/convergence_curve_test.py | 618,619 | 560,674 | test_convert_no_restart | RestartingCurveConverterTest | 619,674 | 619,674 | class RestartingCurveConverterTest(absltest.TestCase):
def test_convert_with_restart(self):
def converter_factory():
return convergence.HypervolumeCurveConverter(
[
pyvizier.MetricInformation(
name='max', goal=pyvizier.ObjectiveMetricGoal.MAXIMIZE
),
... | class RestartingCurveConverterTest(absltest.TestCase):
def test_convert_with_restart(self):
def converter_factory():
return convergence.HypervolumeCurveConverter(
[
pyvizier.MetricInformation(
name='max', goal=pyvizier.ObjectiveMetricGoal.MAXIMIZE
),
... | def converter_factory():
return convergence.HypervolumeCurveConverter(
[
pyvizier.MetricInformation(
name='max', goal=pyvizier.ObjectiveMetricGoal.MAXIMIZE
),
pyvizier.MetricInformation(
name='min', goal=pyvizier.Objective... | def converter_factory():
return convergence.HypervolumeCurveConverter(
[
pyvizier.MetricInformation(
name='max', goal=pyvizier.ObjectiveMetricGoal.MAXIMIZE
),
pyvizier.MetricInformation(
name='min', goal=pyvizier.Objective... | 91 | 669 | 641 | 1,310 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,994 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/analyzers/convergence_curve_test.py | 679,680 | 674,746 | setUp | WinRateConvergenceCurveComparatorAllComparisonsModeTest | 680,700 | 680,700 | class WinRateConvergenceCurveComparatorAllComparisonsModeTest(
absltest.TestCase
):
"""Test the WinRateConvergenceCurveComparator with "all comparisons" mode."""
def setUp(self):
FILL_FUNC_BODY def test_higher_quantile_curve(self):
baseline_length = len(self._baseline_curve.xs)
median_score = converge... | class WinRateConvergenceCurveComparatorAllComparisonsModeTest(
absltest.TestCase
):
"""Test the WinRateConvergenceCurveComparator with "all comparisons" mode."""
def setUp(self):
FILL_FUNC_BODY def test_higher_quantile_curve(self):
baseline_length = len(self._baseline_curve.xs)
median_score = converge... | super(WinRateConvergenceCurveComparatorAllComparisonsModeTest, self).setUp()
xs = np.array(range(0, 20))
xs_t = xs.reshape(1, len(xs))
self._baseline_curve = convergence.ConvergenceCurve(
xs=xs,
ys=np.exp(np.array([-0.9, -1.0, -1.1]).reshape(3, 1) * xs_t),
trend=convergence.Conve... | super(WinRateConvergenceCurveComparatorAllComparisonsModeTest, self).setUp()
xs = np.array(range(0, 20))
xs_t = xs.reshape(1, len(xs))
self._baseline_curve = convergence.ConvergenceCurve(
xs=xs,
ys=np.exp(np.array([-0.9, -1.0, -1.1]).reshape(3, 1) * xs_t),
trend=convergence.Conve... | 43 | 667 | 340 | 1,007 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,995 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/analyzers/convergence_curve_test.py | 892,910 | 887,924 | test_score_one_curve_above_other | WinRateConvergenceCurveComparatorQuantilesModeTest | 910,924 | 910,924 | class WinRateConvergenceCurveComparatorQuantilesModeTest(
parameterized.TestCase
):
"""Tests for WinRateConvergenceCurveComparator with quantiles mode."""
@parameterized.parameters(
{
'ys1': np.array([[11, 12, 20]]),
'ys2': np.array([[1, 2, 10]]),
'res': -0.5,
},
... | class WinRateConvergenceCurveComparatorQuantilesModeTest(
parameterized.TestCase
):
"""Tests for WinRateConvergenceCurveComparator with quantiles mode."""
@parameterized.parameters(
{
'ys1': np.array([[11, 12, 20]]),
'ys2': np.array([[1, 2, 10]]),
'res': -0.5,
},
... | xs1 = np.arange(ys1.shape[1])
xs2 = np.arange(ys2.shape[1])
curve1 = convergence.ConvergenceCurve(
xs=xs1, ys=ys1, trend=convergence.ConvergenceCurve.YTrend.INCREASING
)
curve2 = convergence.ConvergenceCurve(
xs=xs2, ys=ys2, trend=convergence.ConvergenceCurve.YTrend.INCREASING
)
... | xs1 = np.arange(ys1.shape[1])
xs2 = np.arange(ys2.shape[1])
curve1 = convergence.ConvergenceCurve(
xs=xs1, ys=ys1, trend=convergence.ConvergenceCurve.YTrend.INCREASING
)
curve2 = convergence.ConvergenceCurve(
xs=xs2, ys=ys2, trend=convergence.ConvergenceCurve.YTrend.INCREASING
)
... | 29 | 278 | 203 | 481 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,997 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/analyzers/exploration_score_utils_test.py | 198,199 | 82,222 | test_compute_average_marginal_parameter_entropy | ExplorationScoreUtilsTest | 199,222 | 199,222 | class ExplorationScoreUtilsTest(parameterized.TestCase):
@parameterized.named_parameters(
('Continuous', vz.ParameterType.DOUBLE),
('Discrete', vz.ParameterType.DISCRETE),
('Integer', vz.ParameterType.INTEGER),
('Categorical', vz.ParameterType.CATEGORICAL),
)
def test_compute_parameter_en... | class ExplorationScoreUtilsTest(parameterized.TestCase):
@parameterized.named_parameters(
('Continuous', vz.ParameterType.DOUBLE),
('Discrete', vz.ParameterType.DISCRETE),
('Integer', vz.ParameterType.INTEGER),
('Categorical', vz.ParameterType.CATEGORICAL),
)
def test_compute_parameter_en... | min_ent_study, max_ent_study = _generate_min_and_max_ent_studies()
max_ent = (
exploration_score_utils.compute_average_marginal_parameter_entropy(
{'spec_gen': {'hash_': {0: max_ent_study}}}
)
)
min_ent = (
exploration_score_utils.compute_average_marginal_parameter_en... | min_ent_study, max_ent_study = _generate_min_and_max_ent_studies()
max_ent = (
exploration_score_utils.compute_average_marginal_parameter_entropy(
{'spec_gen': {'hash_': {0: max_ent_study}}}
)
)
min_ent = (
exploration_score_utils.compute_average_marginal_parameter_en... | 109 | 1,458 | 336 | 1,794 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,998 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/analyzers/simple_regret_score_test.py | 89,90 | 27,105 | test_score_decreases_with_samples_two_sample | SimpleRegretScoreTest | 90,105 | 90,105 | class SimpleRegretScoreTest(parameterized.TestCase):
# @parameterized.product(
# create_problem=[
# create_continuous_problem,
# create_categorical_problem,
# create_mix_problem,
# ],
# n_features=list(range(10, 20)),
# )
@parameterized.parameters(
... | class SimpleRegretScoreTest(parameterized.TestCase):
# @parameterized.product(
# create_problem=[
# create_continuous_problem,
# create_categorical_problem,
# create_mix_problem,
# ],
# n_features=list(range(10, 20)),
# )
@parameterized.parameters(
... | goal = vz.ObjectiveMetricGoal.MAXIMIZE
baseline_simple_regrets = np.ones(500) + 0.1 * np.random.normal(size=(500,))
candidate_simple_regrets = 1.2 * np.ones(200)
p_value500 = simple_regret_score.t_test_mean_score(
baseline_simple_regrets, candidate_simple_regrets, goal)
baseline_simple_regr... | goal = vz.ObjectiveMetricGoal.MAXIMIZE
baseline_simple_regrets = np.ones(500) + 0.1 * np.random.normal(size=(500,))
candidate_simple_regrets = 1.2 * np.ones(200)
p_value500 = simple_regret_score.t_test_mean_score(
baseline_simple_regrets, candidate_simple_regrets, goal)
baseline_simple_regr... | 51 | 715 | 272 | 987 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
46,999 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/analyzers/state_analyzer_test.py | 86,87 | 27,183 | test_record_conversion | StateAnalyzerTest | 87,120 | 87,120 | class StateAnalyzerTest(absltest.TestCase):
def test_curve_conversion(self):
dim = 10
experimenter = experimenters.BBOBExperimenterFactory('Sphere', dim)()
def _designer_factory(config: vz.ProblemStatement, seed: int):
return random.RandomDesigner(config.search_space, seed=seed)
benchmark_sta... | class StateAnalyzerTest(absltest.TestCase):
def test_curve_conversion(self):
dim = 10
experimenter = experimenters.BBOBExperimenterFactory('Sphere', dim)()
def _designer_factory(config: vz.ProblemStatement, seed: int):
return random.RandomDesigner(config.search_space, seed=seed)
benchmark_sta... | dim = 10
factory = experimenters.BBOBExperimenterFactory('Sphere', dim)
experimenter = factory()
def _designer_factory(config: vz.ProblemStatement, seed: int):
return random.RandomDesigner(config.search_space, seed=seed)
benchmark_state_factory = vzb.DesignerBenchmarkStateFactory(
de... | dim = 10
factory = experimenters.BBOBExperimenterFactory('Sphere', dim)
experimenter = factory()
def _designer_factory(config: vz.ProblemStatement, seed: int):
return random.RandomDesigner(config.search_space, seed=seed)
benchmark_state_factory = vzb.DesignerBenchmarkStateFactory(
de... | 68 | 1,327 | 432 | 1,759 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,000 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/atari100k_experimenter_test.py | 29,33 | 27,65 | test_e2e_evaluation | Atari100KTest | 33,65 | 33,65 | class Atari100KTest(parameterized.TestCase):
@absltest.skip("ALE ROMS must be installed manually.")
@parameterized.parameters('DER', 'DrQ', 'DrQ_eps', 'OTRainbow')
@absltest.skip('Jax versioning not updated in Dopamine.')
def test_e2e_evaluation(self, agent_name):
FILL_FUNC_BODY | class Atari100KTest(parameterized.TestCase):
@absltest.skip("ALE ROMS must be installed manually.")
@parameterized.parameters('DER', 'DrQ', 'DrQ_eps', 'OTRainbow')
@absltest.skip('Jax versioning not updated in Dopamine.')
def test_e2e_evaluation(self, agent_name):
FILL_FUNC_BODY | from vizier._src.benchmarks.experimenters import atari100k_experimenter # pylint: disable=g-import-not-at-top
initial_gin_bindings = {
'Runner.training_steps': 2,
'MaxEpisodeEvalRunner.num_eval_episodes': 2,
'Runner.num_iterations': 2,
'Runner.max_steps_per_episode': 2,
... | from vizier._src.benchmarks.experimenters import atari100k_experimenter # pylint: disable=g-import-not-at-top
initial_gin_bindings = {
'Runner.training_steps': 2,
'MaxEpisodeEvalRunner.num_eval_episodes': 2,
'Runner.num_iterations': 2,
'Runner.max_steps_per_episode': 2,
... | 71 | 110 | 441 | 551 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,001 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/combo/common.py | 97,102 | 1,117 | ising_dense | 102,117 | 102,117 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... | diff_horizontal = interaction_original[0] - interaction_sparsified[0]
diff_vertical = interaction_original[1] - interaction_sparsified[1]
kld = 0
n_spin = covariance.shape[0]
for i in range(n_spin):
i_h, i_v = int(i / ising_grid_h), int(i % ising_grid_h)
for j in range(i, n_spin):
j_h, j_v = in... | diff_horizontal = interaction_original[0] - interaction_sparsified[0]
diff_vertical = interaction_original[1] - interaction_sparsified[1]
kld = 0
n_spin = covariance.shape[0]
for i in range(n_spin):
i_h, i_v = int(i / ising_grid_h), int(i % ising_grid_h)
for j in range(i, n_spin):
j_h, j_v = in... | 87 | 1,560 | 293 | 1,853 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... | |
47,002 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/combo_experimenter_test.py | 33,46 | 30,66 | test_experimenters | ComboExperimenterTest | 47,66 | 46,66 | class ComboExperimenterTest(parameterized.TestCase):
"""Default lamda values found in README of https://github.com/QUVA-Lab/COMBO."""
@parameterized.named_parameters(
('contamination',
functools.partial(
combo_experimenter.ContaminationExperimenter,
lamda=0.01), 20.0, 30.0),
... | class ComboExperimenterTest(parameterized.TestCase):
"""Default lamda values found in README of https://github.com/QUVA-Lab/COMBO."""
@parameterized.named_parameters(
('contamination',
functools.partial(
combo_experimenter.ContaminationExperimenter,
lamda=0.01), 20.0, 30.0),
... | experimenter = experimenter_class()
problem_statement = experimenter.problem_statement()
logging.info(problem_statement)
designer = random.RandomDesigner(
search_space=problem_statement.search_space, seed=None)
suggestions = designer.suggest(5)
trials = [suggestion.to_trial() for sugge... | """Tests if entire pipeline works and objective values make sense."""
experimenter = experimenter_class()
problem_statement = experimenter.problem_statement()
logging.info(problem_statement)
designer = random.RandomDesigner(
search_space=problem_statement.search_space, seed=None)
sugge... | 43 | 267 | 251 | 518 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,003 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/discretizing_experimenter.py | 129,136 | 28,209 | create_with_grid | DiscretizingExperimenter | 155,209 | 136,209 | class DiscretizingExperimenter(experimenter.Experimenter):
"""DiscretizingExperimenter discretizes the parameters of search space."""
def __init__(
self,
exptr: experimenter.Experimenter,
discretization: Mapping[str, pyvizier.MonotypeParameterSequence],
*,
allow_oov: bool = False,
)... | class DiscretizingExperimenter(experimenter.Experimenter):
"""DiscretizingExperimenter discretizes the parameters of search space."""
def __init__(
self,
exptr: experimenter.Experimenter,
discretization: Mapping[str, pyvizier.MonotypeParameterSequence],
*,
allow_oov: bool = False,
)... | if isinstance(convert_to_str, bool):
convert_to_str = {
k: convert_to_str for k in grid_discretization_count.keys()
}
if set(grid_discretization_count.keys()) != set(convert_to_str.keys()):
raise ValueError(
f'Grid discretization keys {grid_discretization_count.keys()} mus... | """Creates Experimenter with continuous parameters discretized via grid.
Note that the grid is generated according to ModelInputConverter scaling.
Args:
exptr: Experimenter with continuous parameters to be discretized.
grid_discretization_count: Mapping from parameter names to number of
... | 169 | 1,349 | 678 | 2,027 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,004 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/discretizing_experimenter_test.py | 29,44 | 27,159 | testNumpyExperimenter | DiscretizingExperimenterTest | 44,84 | 44,84 | class DiscretizingExperimenterTest(parameterized.TestCase):
@parameterized.named_parameters(
('Sphere', bbob.Sphere), ('Rastrigin', bbob.Rastrigin),
('BuecheRastrigin', bbob.BuecheRastrigin),
('LinearSlope', bbob.LinearSlope),
('AttractiveSector', bbob.AttractiveSector),
('StepEllipsoid... | class DiscretizingExperimenterTest(parameterized.TestCase):
@parameterized.named_parameters(
('Sphere', bbob.Sphere), ('Rastrigin', bbob.Rastrigin),
('BuecheRastrigin', bbob.BuecheRastrigin),
('LinearSlope', bbob.LinearSlope),
('AttractiveSector', bbob.AttractiveSector),
('StepEllipsoid... | dim = 3
exptr = numpy_experimenter.NumpyExperimenter(
func, bbob.DefaultBBOBProblemStatement(dim))
# Asserts parameters are the same.
parameters = list(exptr.problem_statement().search_space.parameters)
self.assertLen(parameters, dim)
discretization = {
parameters[0].name: ['-1... | dim = 3
exptr = numpy_experimenter.NumpyExperimenter(
func, bbob.DefaultBBOBProblemStatement(dim))
# Asserts parameters are the same.
parameters = list(exptr.problem_statement().search_space.parameters)
self.assertLen(parameters, dim)
discretization = {
parameters[0].name: ['-1... | 79 | 1,248 | 495 | 1,743 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,005 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/experimenter_factory.py | 101,103 | 70,107 | recover | BBOBExperimenterFactory | 103,107 | 103,107 | @attr.define
class BBOBExperimenterFactory(SerializableExperimenterFactory):
"""Factory for a BBOB function."""
# Should be a BBOB function name in bbob.py (name should match exactly).
name: str = attr.field(default='', validator=attr.validators.instance_of(str))
dim: int = attr.field(
default=1,
v... | @attr.define
class BBOBExperimenterFactory(SerializableExperimenterFactory):
"""Factory for a BBOB function."""
# Should be a BBOB function name in bbob.py (name should match exactly).
name: str = attr.field(default='', validator=attr.validators.instance_of(str))
dim: int = attr.field(
default=1,
v... | metadata_dict = json.loads(metadata[BBOB_FACTORY_KEY])
return cls(**metadata_dict)
| metadata_dict = json.loads(metadata[BBOB_FACTORY_KEY])
return cls(**metadata_dict)
| 5 | 438 | 40 | 478 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,006 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/experimenter_factory.py | 220,224 | 107,248 | recover | SingleObjectiveExperimenterFactory | 224,248 | 224,248 | @attr.define
class SingleObjectiveExperimenterFactory(SerializableExperimenterFactory):
"""Factory for a single objective Experimenter.
Attributes:
base_factory:
shift: An array of doubles that is broadcastable to dim of search space.
should_restrict: Whether to restrict the parameter bounds of search ... | @attr.define
class SingleObjectiveExperimenterFactory(SerializableExperimenterFactory):
"""Factory for a single objective Experimenter.
Attributes:
base_factory:
shift: An array of doubles that is broadcastable to dim of search space.
should_restrict: Whether to restrict the parameter bounds of search ... | if BBOB_FACTORY_KEY in metadata:
base_factory = BBOBExperimenterFactory.recover(metadata)
else:
raise serializable.DecodeError(
f'No valid base factory found in {metadata}'
)
metadata_dict = json.loads(
metadata[SINGLE_OBJECTIVE_FACTORY_KEY], cls=json_utils.NumpyDecoder
... | if BBOB_FACTORY_KEY in metadata:
base_factory = BBOBExperimenterFactory.recover(metadata)
else:
raise serializable.DecodeError(
f'No valid base factory found in {metadata}'
)
metadata_dict = json.loads(
metadata[SINGLE_OBJECTIVE_FACTORY_KEY], cls=json_utils.NumpyDecoder
... | 67 | 1,407 | 293 | 1,700 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,009 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/infeasible_experimenter.py | 41,42 | 30,62 | evaluate | HashingInfeasibleExperimenter | 42,52 | 42,52 | @attrs.define
class HashingInfeasibleExperimenter(experimenter.Experimenter):
"""Simulates randomly selected (deterministic) infeasibility function."""
exptr: experimenter.Experimenter = attrs.field()
infeasible_prob: float = attrs.field(default=0.2, kw_only=True)
seed: int = attrs.field(default=0, kw_only=Tru... | @attrs.define
class HashingInfeasibleExperimenter(experimenter.Experimenter):
"""Simulates randomly selected (deterministic) infeasibility function."""
exptr: experimenter.Experimenter = attrs.field()
infeasible_prob: float = attrs.field(default=0.2, kw_only=True)
seed: int = attrs.field(default=0, kw_only=Tru... | metric_configs = self._problem.metric_information
for suggestion in suggestions:
if self._is_infeasible(suggestion.parameters):
suggestion.complete(
vz.Measurement(metrics={mc.name: np.nan for mc in metric_configs}),
infeasibility_reason='HashingInfeasibleExperimenter',
... | metric_configs = self._problem.metric_information
for suggestion in suggestions:
if self._is_infeasible(suggestion.parameters):
suggestion.complete(
vz.Measurement(metrics={mc.name: np.nan for mc in metric_configs}),
infeasibility_reason='HashingInfeasibleExperimenter',
... | 20 | 292 | 122 | 414 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,010 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/infeasible_experimenter.py | 82,83 | 62,99 | evaluate | ParamRegionInfeasibleExperimenter | 83,97 | 83,97 | @attrs.define
class ParamRegionInfeasibleExperimenter(experimenter.Experimenter):
"""Selects a parameter and splits its values into feasible/infeasible."""
exptr: experimenter.Experimenter = attrs.field()
parameter_name: str = attrs.field()
infeasible_interval: tuple[float, float] = attrs.field(
default=... | @attrs.define
class ParamRegionInfeasibleExperimenter(experimenter.Experimenter):
"""Selects a parameter and splits its values into feasible/infeasible."""
exptr: experimenter.Experimenter = attrs.field()
parameter_name: str = attrs.field()
infeasible_interval: tuple[float, float] = attrs.field(
default=... | param_features = self._converter.convert(suggestions)
metric_configs = self._problem.metric_information
for sugg_index, param_feature in enumerate(param_features):
p = param_feature.item()
suggestion = suggestions[sugg_index]
if self.infeasible_interval[0] <= p <= self.infeasible_interval... | param_features = self._converter.convert(suggestions)
metric_configs = self._problem.metric_information
for sugg_index, param_feature in enumerate(param_features):
p = param_feature.item()
suggestion = suggestions[sugg_index]
if self.infeasible_interval[0] <= p <= self.infeasible_interval... | 34 | 309 | 186 | 495 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,011 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/infeasible_experimenter_test.py | 26,27 | 24,46 | test_consistency | HashingInfeasibleExperimenterTest | 27,46 | 27,46 | class HashingInfeasibleExperimenterTest(absltest.TestCase):
def test_consistency(self):
FILL_FUNC_BODY | class HashingInfeasibleExperimenterTest(absltest.TestCase):
def test_consistency(self):
FILL_FUNC_BODY | exptr = numpy_experimenter.NumpyExperimenter(
bbob.Sphere, bbob.DefaultBBOBProblemStatement(2)
)
exptr = infeasible_experimenter.HashingInfeasibleExperimenter(
exptr, infeasible_prob=0.5, seed=0
)
for i in range(10):
trials = [vz.Trial(parameters={'x0': i, 'x1': -i}) for _ in ... | exptr = numpy_experimenter.NumpyExperimenter(
bbob.Sphere, bbob.DefaultBBOBProblemStatement(2)
)
exptr = infeasible_experimenter.HashingInfeasibleExperimenter(
exptr, infeasible_prob=0.5, seed=0
)
for i in range(10):
trials = [vz.Trial(parameters={'x0': i, 'x1': -i}) for _ in ... | 48 | 40 | 264 | 304 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,012 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/infeasible_experimenter_test.py | 48,49 | 46,64 | test_e2e | ParamRegionInfeasibleExperimenterTest | 49,64 | 49,64 | class ParamRegionInfeasibleExperimenterTest(absltest.TestCase):
def test_e2e(self):
FILL_FUNC_BODY | class ParamRegionInfeasibleExperimenterTest(absltest.TestCase):
def test_e2e(self):
FILL_FUNC_BODY | exptr = numpy_experimenter.NumpyExperimenter(
bbob.Sphere, bbob.DefaultBBOBProblemStatement(2)
)
exptr = infeasible_experimenter.ParamRegionInfeasibleExperimenter(
exptr, parameter_name='x0', infeasible_interval=(0.0, 0.5)
)
infeasible_trial = vz.Trial(parameters={'x0': -3.5, 'x1': ... | exptr = numpy_experimenter.NumpyExperimenter(
bbob.Sphere, bbob.DefaultBBOBProblemStatement(2)
)
exptr = infeasible_experimenter.ParamRegionInfeasibleExperimenter(
exptr, parameter_name='x0', infeasible_interval=(0.0, 0.5)
)
infeasible_trial = vz.Trial(parameters={'x0': -3.5, 'x1': ... | 30 | 42 | 243 | 285 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,013 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/l1_categorical_experimenter.py | 93,94 | 32,96 | problem_statement | L1CategorialExperimenter | 95,96 | 94,96 | class L1CategorialExperimenter(experimenter.Experimenter):
"""Categorical experimenter associated minimum optimization problem."""
def __init__(
self,
*,
num_categories: Sequence[int],
optimum: Optional[Sequence[int]] = None,
seed: Optional[int] = None,
verbose: bool = False,
... | class L1CategorialExperimenter(experimenter.Experimenter):
"""Categorical experimenter associated minimum optimization problem."""
def __init__(
self,
*,
num_categories: Sequence[int],
optimum: Optional[Sequence[int]] = None,
seed: Optional[int] = None,
verbose: bool = False,
... | return copy.deepcopy(self._problem)
| """See superclass."""
return copy.deepcopy(self._problem)
| 2 | 702 | 13 | 715 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,014 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/l1_categorical_experimenter_test.py | 29,30 | 26,76 | test_evaluate_optimum | L1CategoricalExperimenterTest | 30,39 | 30,39 | class L1CategoricalExperimenterTest(parameterized.TestCase):
"""Tests for CategoricalExperimenter."""
def test_evaluate_optimum(self):
FILL_FUNC_BODY def test_evaluate_non_optimum(self):
optimum = [0, 1]
exptr = l1_categorical_experimenter.L1CategorialExperimenter(
num_categories=[2, 2], optimum=o... | class L1CategoricalExperimenterTest(parameterized.TestCase):
"""Tests for CategoricalExperimenter."""
def test_evaluate_optimum(self):
FILL_FUNC_BODY def test_evaluate_non_optimum(self):
optimum = [0, 1]
exptr = l1_categorical_experimenter.L1CategorialExperimenter(
num_categories=[2, 2], optimum=o... | optimum = [0, 1]
exptr = l1_categorical_experimenter.L1CategorialExperimenter(
num_categories=[2, 2], optimum=optimum)
suggestion = vz.Trial(parameters={'c0': '0', 'c1': '1'})
exptr.evaluate([suggestion])
self.assertEqual(
suggestion.final_measurement_or_die.metrics['objective'].valu... | optimum = [0, 1]
exptr = l1_categorical_experimenter.L1CategorialExperimenter(
num_categories=[2, 2], optimum=optimum)
suggestion = vz.Trial(parameters={'c0': '0', 'c1': '1'})
exptr.evaluate([suggestion])
self.assertEqual(
suggestion.final_measurement_or_die.metrics['objective'].valu... | 21 | 643 | 135 | 778 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,015 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/multiobjective_experimenter.py | 78,79 | 27,101 | evaluate | MultiObjectiveExperimenter | 79,96 | 79,96 | class MultiObjectiveExperimenter(experimenter.Experimenter):
"""MultiObjective Experimenter from a Dict of Experimenters."""
def __init__(
self,
exptrs: Dict[str, experimenter.Experimenter],
):
"""Creates a MultiObjectiveExperimenter from a Dict of Experimenters.
Specifically, the Experiment... | class MultiObjectiveExperimenter(experimenter.Experimenter):
"""MultiObjective Experimenter from a Dict of Experimenters."""
def __init__(
self,
exptrs: Dict[str, experimenter.Experimenter],
):
"""Creates a MultiObjectiveExperimenter from a Dict of Experimenters.
Specifically, the Experiment... | suggestions_copy = copy.deepcopy(suggestions)
measurements = [pyvizier.Measurement() for _ in suggestions]
for name, exptr in self._exptrs.items():
exptr.evaluate(suggestions_copy)
previous_name = self._previous_names[name]
for idx, copied in enumerate(suggestions_copy):
measuremen... | suggestions_copy = copy.deepcopy(suggestions)
measurements = [pyvizier.Measurement() for _ in suggestions]
for name, exptr in self._exptrs.items():
exptr.evaluate(suggestions_copy)
previous_name = self._previous_names[name]
for idx, copied in enumerate(suggestions_copy):
measuremen... | 46 | 635 | 185 | 820 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,016 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/multiobjective_experimenter_test.py | 61,62 | 26,74 | test_dimension_mismatch | MultiobjectiveExperimenterTest | 62,74 | 62,74 | class MultiobjectiveExperimenterTest(absltest.TestCase):
def test_mulitobjective_numpy(self):
dim = 2
func1 = bbob.Sphere
func2 = bbob.Rastrigin
exptr1 = numpy_experimenter.NumpyExperimenter(
func1, bbob.DefaultBBOBProblemStatement(dim)
)
exptr2 = numpy_experimenter.NumpyExperimenter(... | class MultiobjectiveExperimenterTest(absltest.TestCase):
def test_mulitobjective_numpy(self):
dim = 2
func1 = bbob.Sphere
func2 = bbob.Rastrigin
exptr1 = numpy_experimenter.NumpyExperimenter(
func1, bbob.DefaultBBOBProblemStatement(dim)
)
exptr2 = numpy_experimenter.NumpyExperimenter(... | exptr1 = numpy_experimenter.NumpyExperimenter(
bbob.Sphere, bbob.DefaultBBOBProblemStatement(dimension=3)
)
exptr2 = numpy_experimenter.NumpyExperimenter(
bbob.Rastrigin, bbob.DefaultBBOBProblemStatement(dimension=6)
)
with self.assertRaisesRegex(ValueError, 'space must match'):
... | exptr1 = numpy_experimenter.NumpyExperimenter(
bbob.Sphere, bbob.DefaultBBOBProblemStatement(dimension=3)
)
exptr2 = numpy_experimenter.NumpyExperimenter(
bbob.Rastrigin, bbob.DefaultBBOBProblemStatement(dimension=6)
)
with self.assertRaisesRegex(ValueError, 'space must match'):
... | 23 | 423 | 172 | 595 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,018 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/noisy_experimenter.py | 58,59 | 28,74 | evaluate | NoisyExperimenter | 59,70 | 59,70 | @attr.define(auto_attribs=True)
class NoisyExperimenter(experimenter.Experimenter):
"""NoisyExperimenter applies noise to all metric(s) in final measurement.
It stores the unnoised metric as metric_name + '_before_noise'. The noise
function is a callable that is applied to metrics.
"""
exptr: experimenter.E... | @attr.define(auto_attribs=True)
class NoisyExperimenter(experimenter.Experimenter):
"""NoisyExperimenter applies noise to all metric(s) in final measurement.
It stores the unnoised metric as metric_name + '_before_noise'. The noise
function is a callable that is applied to metrics.
"""
exptr: experimenter.E... | self.exptr.evaluate(suggestions)
for suggestion in suggestions:
if suggestion.final_measurement is None:
continue
metric_dict_with_noise = {}
for name, metric in suggestion.final_measurement.metrics.items():
metric_dict_with_noise[name] = pyvizier.Metric(
value=self... | self.exptr.evaluate(suggestions)
for suggestion in suggestions:
if suggestion.final_measurement is None:
continue
metric_dict_with_noise = {}
for name, metric in suggestion.final_measurement.metrics.items():
metric_dict_with_noise[name] = pyvizier.Metric(
value=self... | 30 | 369 | 150 | 519 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,019 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/noisy_experimenter_test.py | 72,85 | 30,155 | testGaussianNoiseApply | NoisyExperimenterTest | 85,113 | 85,113 | class NoisyExperimenterTest(parameterized.TestCase):
@parameterized.named_parameters(
('Sphere', bbob.Sphere), ('Rastrigin', bbob.Rastrigin),
('BuecheRastrigin', bbob.BuecheRastrigin),
('LinearSlope', bbob.LinearSlope),
('AttractiveSector', bbob.AttractiveSector),
('StepEllipsoidal', bb... | class NoisyExperimenterTest(parameterized.TestCase):
@parameterized.named_parameters(
('Sphere', bbob.Sphere), ('Rastrigin', bbob.Rastrigin),
('BuecheRastrigin', bbob.BuecheRastrigin),
('LinearSlope', bbob.LinearSlope),
('AttractiveSector', bbob.AttractiveSector),
('StepEllipsoidal', bb... | dim = 2
exptr = numpy_experimenter.NumpyExperimenter(
bbob.Sphere, bbob.DefaultBBOBProblemStatement(dim))
noisy_exptr = noisy_experimenter.NoisyExperimenter.from_type(
exptr=exptr, noise_type=noise
)
parameters = exptr.problem_statement().search_space.parameters
t = pyvizier.Tri... | dim = 2
exptr = numpy_experimenter.NumpyExperimenter(
bbob.Sphere, bbob.DefaultBBOBProblemStatement(dim))
noisy_exptr = noisy_experimenter.NoisyExperimenter.from_type(
exptr=exptr, noise_type=noise
)
parameters = exptr.problem_statement().search_space.parameters
t = pyvizier.Tri... | 67 | 1,477 | 415 | 1,892 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,020 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/normalizing_experimenter.py | 85,86 | 28,102 | evaluate | NormalizingExperimenter | 86,97 | 86,97 | class NormalizingExperimenter(experimenter.Experimenter):
"""Normalizes an Experimenter output via dividing by L1 norm at grid points."""
def __init__(
self,
exptr: experimenter.Experimenter,
num_normalization_samples: int = 100,
):
"""Normalizing experimenter uses a grid to estimate a norm... | class NormalizingExperimenter(experimenter.Experimenter):
"""Normalizes an Experimenter output via dividing by L1 norm at grid points."""
def __init__(
self,
exptr: experimenter.Experimenter,
num_normalization_samples: int = 100,
):
"""Normalizing experimenter uses a grid to estimate a norm... | self._exptr.evaluate(suggestions)
for suggestion in suggestions:
if suggestion.final_measurement is None:
continue
normalized_metrics: Dict[str, pyvizier.Metric] = {}
for name, metric in suggestion.final_measurement.metrics.items():
normalized_metrics[name] = pyvizier.Metric(
... | self._exptr.evaluate(suggestions)
for suggestion in suggestions:
if suggestion.final_measurement is None:
continue
normalized_metrics: Dict[str, pyvizier.Metric] = {}
for name, metric in suggestion.final_measurement.metrics.items():
normalized_metrics[name] = pyvizier.Metric(
... | 30 | 708 | 140 | 848 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,021 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/normalizing_experimenter_test.py | 28,50 | 26,75 | testNormalizationApply | NormalizingExperimenterTest | 50,75 | 50,75 | class NormalizingExperimenterTest(parameterized.TestCase):
@parameterized.named_parameters(
('Sphere', bbob.Sphere),
('Rastrigin', bbob.Rastrigin),
('BuecheRastrigin', bbob.BuecheRastrigin),
('LinearSlope', bbob.LinearSlope),
('AttractiveSector', bbob.AttractiveSector),
('StepElli... | class NormalizingExperimenterTest(parameterized.TestCase):
@parameterized.named_parameters(
('Sphere', bbob.Sphere),
('Rastrigin', bbob.Rastrigin),
('BuecheRastrigin', bbob.BuecheRastrigin),
('LinearSlope', bbob.LinearSlope),
('AttractiveSector', bbob.AttractiveSector),
('StepElli... | dim = 5
exptr = numpy_experimenter.NumpyExperimenter(
func, bbob.DefaultBBOBProblemStatement(dim)
)
normalizing_exptr = normalizing_experimenter.NormalizingExperimenter(
exptr=exptr
)
parameters = exptr.problem_statement().search_space.parameters
self.assertLen(parameters, d... | dim = 5
exptr = numpy_experimenter.NumpyExperimenter(
func, bbob.DefaultBBOBProblemStatement(dim)
)
normalizing_exptr = normalizing_experimenter.NormalizingExperimenter(
exptr=exptr
)
parameters = exptr.problem_statement().search_space.parameters
self.assertLen(parameters, d... | 43 | 407 | 266 | 673 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,022 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/numpy_experimenter.py | 30,31 | 1,117 | _get_name | 32,40 | 31,40 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... | if hasattr(f, '__name__'):
return f.__name__
# Next clause handles functools.partial objects.
if hasattr(f, 'func') and hasattr(f.func, '__name__'):
return f.func.__name__
return repr(f)
| """Gets the name of underlying objects."""
if hasattr(f, '__name__'):
return f.__name__
# Next clause handles functools.partial objects.
if hasattr(f, 'func') and hasattr(f.func, '__name__'):
return f.func.__name__
return repr(f)
| 21 | 1,088 | 86 | 1,174 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... | |
47,023 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/numpy_experimenter.py | 97,98 | 40,117 | evaluate | NumpyExperimenter | 98,115 | 98,115 | class NumpyExperimenter(experimenter.Experimenter):
"""NumpyExperimenters take a deterministic function on ndarrays."""
def __init__(
self,
impl: Callable[[np.ndarray], float],
problem_statement: pyvizier.ProblemStatement,
):
"""NumpyExperimenter with analytic function impl for one metric.
... | class NumpyExperimenter(experimenter.Experimenter):
"""NumpyExperimenters take a deterministic function on ndarrays."""
def __init__(
self,
impl: Callable[[np.ndarray], float],
problem_statement: pyvizier.ProblemStatement,
):
"""NumpyExperimenter with analytic function impl for one metric.
... | # Features has shape (num_trials, num_features).
features = self._converter.to_features(suggestions)
for idx, suggestion in enumerate(suggestions):
val = self.impl(features[idx])
if math.isfinite(val):
suggestion.complete(
pyvizier.Measurement(metrics={self._metric_name: val}... | # Features has shape (num_trials, num_features).
features = self._converter.to_features(suggestions)
for idx, suggestion in enumerate(suggestions):
val = self.impl(features[idx])
if math.isfinite(val):
suggestion.complete(
pyvizier.Measurement(metrics={self._metric_name: val}... | 38 | 646 | 188 | 834 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,024 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/numpy_experimenter_test.py | 30,45 | 28,97 | testNumpyExperimenter | NumpyExperimenterTest | 45,64 | 45,64 | class NumpyExperimenterTest(parameterized.TestCase):
@parameterized.named_parameters(
('Sphere', bbob.Sphere), ('Rastrigin', bbob.Rastrigin),
('BuecheRastrigin', bbob.BuecheRastrigin),
('LinearSlope', bbob.LinearSlope),
('AttractiveSector', bbob.AttractiveSector),
('StepEllipsoidal', bb... | class NumpyExperimenterTest(parameterized.TestCase):
@parameterized.named_parameters(
('Sphere', bbob.Sphere), ('Rastrigin', bbob.Rastrigin),
('BuecheRastrigin', bbob.BuecheRastrigin),
('LinearSlope', bbob.LinearSlope),
('AttractiveSector', bbob.AttractiveSector),
('StepEllipsoidal', bb... | dim = 2
exptr = numpy_experimenter.NumpyExperimenter(
func, bbob.DefaultBBOBProblemStatement(dim))
parameters = exptr.problem_statement().search_space.parameters
self.assertLen(parameters, dim)
t = pyvizier.Trial(parameters={
param.name: float(index) for index, param in enumerate(p... | dim = 2
exptr = numpy_experimenter.NumpyExperimenter(
func, bbob.DefaultBBOBProblemStatement(dim))
parameters = exptr.problem_statement().search_space.parameters
self.assertLen(parameters, dim)
t = pyvizier.Trial(parameters={
param.name: float(index) for index, param in enumerate(p... | 35 | 747 | 235 | 982 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,025 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/permuting_experimenter.py | 85,86 | 28,101 | _permute | PermutingExperimenter | 87,99 | 86,99 | class PermutingExperimenter(experimenter.Experimenter):
"""PermutingExperimenter permutes discrete/categorical parameters."""
def __init__(
self,
exptr: experimenter.Experimenter,
parameters_to_permute: Sequence[str],
seed: int = 0,
):
"""PermutingExperiment permutes discrete paramete... | class PermutingExperimenter(experimenter.Experimenter):
"""PermutingExperimenter permutes discrete/categorical parameters."""
def __init__(
self,
exptr: experimenter.Experimenter,
parameters_to_permute: Sequence[str],
seed: int = 0,
):
"""PermutingExperiment permutes discrete paramete... | for suggestion in suggestions:
new_parameters = {}
for name, parameter in suggestion.parameters.items():
if name in self._parameter_permutation_dict:
permutation_dict = self._parameter_permutation_dict[name]
print('permutation dict', permutation_dict)
logging.info('... | """Permutes parameter values in place."""
for suggestion in suggestions:
new_parameters = {}
for name, parameter in suggestion.parameters.items():
if name in self._parameter_permutation_dict:
permutation_dict = self._parameter_permutation_dict[name]
print('permutation dic... | 36 | 687 | 146 | 833 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,026 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/permuting_experimenter_test.py | 30,52 | 28,99 | testNumpyExperimenter | PermutingExperimenterTest | 52,99 | 52,99 | class PermutingExperimenterTest(parameterized.TestCase):
@parameterized.named_parameters(
('Sphere', bbob.Sphere),
('Rastrigin', bbob.Rastrigin),
('BuecheRastrigin', bbob.BuecheRastrigin),
('LinearSlope', bbob.LinearSlope),
('AttractiveSector', bbob.AttractiveSector),
('StepEllips... | class PermutingExperimenterTest(parameterized.TestCase):
@parameterized.named_parameters(
('Sphere', bbob.Sphere),
('Rastrigin', bbob.Rastrigin),
('BuecheRastrigin', bbob.BuecheRastrigin),
('LinearSlope', bbob.LinearSlope),
('AttractiveSector', bbob.AttractiveSector),
('StepEllips... | dim = 3
exptr = numpy_experimenter.NumpyExperimenter(
func, bbob.DefaultBBOBProblemStatement(dim)
)
# Asserts parameters are the same.
parameters = list(exptr.problem_statement().search_space.parameters)
self.assertLen(parameters, dim)
discretization = {
parameters[0].name:... | dim = 3
exptr = numpy_experimenter.NumpyExperimenter(
func, bbob.DefaultBBOBProblemStatement(dim)
)
# Asserts parameters are the same.
parameters = list(exptr.problem_statement().search_space.parameters)
self.assertLen(parameters, dim)
discretization = {
parameters[0].name:... | 86 | 411 | 540 | 951 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,028 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/sign_flip_experimenter.py | 55,56 | 23,70 | evaluate | SignFlipExperimenter | 56,70 | 56,70 | class SignFlipExperimenter(experimenter_lib.Experimenter):
"""Flips the goal and objective values of the wrapped Experimenter."""
def __init__(
self,
experimenter: experimenter_lib.Experimenter,
flip_objectives_only: bool = True,
):
"""Init.
Args:
experimenter: Original experimen... | class SignFlipExperimenter(experimenter_lib.Experimenter):
"""Flips the goal and objective values of the wrapped Experimenter."""
def __init__(
self,
experimenter: experimenter_lib.Experimenter,
flip_objectives_only: bool = True,
):
"""Init.
Args:
experimenter: Original experimen... | self._experimenter.evaluate(suggestions)
for suggestion in suggestions:
if suggestion.final_measurement is None:
continue
metric_dict = {}
for name, metric in suggestion.final_measurement.metrics.items():
if self._flip_objectives_only and name in self._original_objectives:
... | self._experimenter.evaluate(suggestions)
for suggestion in suggestions:
if suggestion.final_measurement is None:
continue
metric_dict = {}
for name, metric in suggestion.final_measurement.metrics.items():
if self._flip_objectives_only and name in self._original_objectives:
... | 44 | 379 | 198 | 577 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,029 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/sign_flip_experimenter_test.py | 33,35 | 31,89 | test_flipped_bbob | SignFlipExperimenterTest | 35,89 | 35,89 | class SignFlipExperimenterTest(parameterized.TestCase):
@parameterized.parameters((True,), (False))
def test_flipped_bbob(self, flip_objectives_only: bool):
FILL_FUNC_BODY | class SignFlipExperimenterTest(parameterized.TestCase):
@parameterized.parameters((True,), (False))
def test_flipped_bbob(self, flip_objectives_only: bool):
FILL_FUNC_BODY | exptr = numpy_experimenter.NumpyExperimenter(
bbob.Sphere, bbob.DefaultBBOBProblemStatement(2)
)
exptr = noisy_experimenter.NoisyExperimenter(
exptr, noise_fn=lambda v: v - 1
)
flipped_exptr = sign_flip_experimenter.SignFlipExperimenter(
exptr, flip_objectives_only
)
... | exptr = numpy_experimenter.NumpyExperimenter(
bbob.Sphere, bbob.DefaultBBOBProblemStatement(2)
)
exptr = noisy_experimenter.NoisyExperimenter(
exptr, noise_fn=lambda v: v - 1
)
flipped_exptr = sign_flip_experimenter.SignFlipExperimenter(
exptr, flip_objectives_only
)
... | 76 | 62 | 606 | 668 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,030 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/sparse_experimenter.py | 99,114 | 43,159 | create | SparseExperimenter | 133,159 | 114,159 | class SparseExperimenter(experimenter.Experimenter):
"""Sparse experimenter.
The sparse experimenter allows expanding an experimenter's search space with
additional placeholder parameters that don't have affect during evaluation.
"""
def __init__(
self,
experiment: experimenter.Experimenter,
... | class SparseExperimenter(experimenter.Experimenter):
"""Sparse experimenter.
The sparse experimenter allows expanding an experimenter's search space with
additional placeholder parameters that don't have affect during evaluation.
"""
def __init__(
self,
experiment: experimenter.Experimenter,
... | sparse_search_space = vz.SearchSpace()
for idx in range(float_count):
sparse_search_space.root.add_float_param(
name='FLOAT' + str(idx),
min_value=float_min_value or _DEFAULT_FLOAT_BOUNDS[0],
max_value=float_max_value or _DEFAULT_FLOAT_BOUNDS[1],
)
for idx in range(... | """Creates a sparse experimenter with different parameter types.
Arguments:
experiment: The experimenter to add sparsity to.
float_count: The number of FLOAT sparse parameters to add.
int_count: The number of INT sparse parameters to add.
discrete_count: The number of DISCRETE sparse pa... | 60 | 981 | 401 | 1,382 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,031 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/sparse_experimenter_test.py | 67,81 | 28,124 | test_sparse_experimenter_create | SparseExperimenterTest | 81,124 | 81,124 | class SparseExperimenterTest(parameterized.TestCase):
@parameterized.parameters(
vz.ParameterConfig.factory('_', bounds=(0.0, 5.0)),
vz.ParameterConfig.factory('_', bounds=(0, 5)),
vz.ParameterConfig.factory('_', feasible_values=[1, 2, 3]),
vz.ParameterConfig.factory('_', feasible_values=['a'... | class SparseExperimenterTest(parameterized.TestCase):
@parameterized.parameters(
vz.ParameterConfig.factory('_', bounds=(0.0, 5.0)),
vz.ParameterConfig.factory('_', bounds=(0, 5)),
vz.ParameterConfig.factory('_', feasible_values=[1, 2, 3]),
vz.ParameterConfig.factory('_', feasible_values=['a'... | sphere_experimenter = experimenter_factory.BBOBExperimenterFactory(
name='Sphere', dim=2
)()
experimenter = sparse_experimenter.SparseExperimenter.create(
experiment=sphere_experimenter,
float_count=float_count,
int_count=int_count,
discrete_count=discrete_count,
... | sphere_experimenter = experimenter_factory.BBOBExperimenterFactory(
name='Sphere', dim=2
)()
experimenter = sparse_experimenter.SparseExperimenter.create(
experiment=sphere_experimenter,
float_count=float_count,
int_count=int_count,
discrete_count=discrete_count,
... | 125 | 731 | 593 | 1,324 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,032 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/switch_experimenter.py | 62,63 | 26,83 | problem_statement | SwitchExperimenter | 63,83 | 63,83 | @attrs.define
class SwitchExperimenter(experimenter.Experimenter):
"""Creates conditional search space from multiple experimenters via a switch."""
experimenters: Sequence[experimenter.Experimenter] = attrs.field()
_switch_param_name: str = attrs.field(default='switch')
_metric_name: str = attrs.field(default... | @attrs.define
class SwitchExperimenter(experimenter.Experimenter):
"""Creates conditional search space from multiple experimenters via a switch."""
experimenters: Sequence[experimenter.Experimenter] = attrs.field()
_switch_param_name: str = attrs.field(default='switch')
_metric_name: str = attrs.field(default... | problem_statement = vz.ProblemStatement()
children = []
for i, problem in enumerate(self._exptr_problems):
for pc in problem.search_space.parameters:
children.append(([i], pc))
switching_param = vz.ParameterConfig.factory(
self._switch_param_name,
feasible_values=range(le... | problem_statement = vz.ProblemStatement()
children = []
for i, problem in enumerate(self._exptr_problems):
for pc in problem.search_space.parameters:
children.append(([i], pc))
switching_param = vz.ParameterConfig.factory(
self._switch_param_name,
feasible_values=range(le... | 33 | 449 | 208 | 657 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,033 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/switch_experimenter_test.py | 26,27 | 24,44 | test_switch_experimenter | SwitchExperimenterTest | 27,44 | 27,44 | class SwitchExperimenterTest(absltest.TestCase):
def test_switch_experimenter(self):
FILL_FUNC_BODY | class SwitchExperimenterTest(absltest.TestCase):
def test_switch_experimenter(self):
FILL_FUNC_BODY | exptr0 = numpy_experimenter.NumpyExperimenter(
lambda x: 0.0 * x, bbob.DefaultBBOBProblemStatement(1, metric_name='0')
)
exptr1 = numpy_experimenter.NumpyExperimenter(
lambda x: 1.0 * x, bbob.DefaultBBOBProblemStatement(1, metric_name='1')
)
switch_exptr = switch_experimenter.SwitchE... | exptr0 = numpy_experimenter.NumpyExperimenter(
lambda x: 0.0 * x, bbob.DefaultBBOBProblemStatement(1, metric_name='0')
)
exptr1 = numpy_experimenter.NumpyExperimenter(
lambda x: 1.0 * x, bbob.DefaultBBOBProblemStatement(1, metric_name='1')
)
switch_exptr = switch_experimenter.SwitchE... | 48 | 38 | 316 | 354 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,034 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/synthetic/bbob_test.py | 31,35 | 29,46 | test_bbob_problem_statement_scale_type | BbobTest | 35,39 | 35,39 | class BbobTest(parameterized.TestCase):
@parameterized.parameters(
{'scale_type': vz.ScaleType.LINEAR}, {'scale_type': vz.ScaleType.LOG}
)
def test_bbob_problem_statement_scale_type(self, scale_type):
FILL_FUNC_BODY def test_rotation(self):
rotation = bbob._R(5, sys.maxsize, 'A', b'XAFDAF', 3)
rot... | class BbobTest(parameterized.TestCase):
@parameterized.parameters(
{'scale_type': vz.ScaleType.LINEAR}, {'scale_type': vz.ScaleType.LOG}
)
def test_bbob_problem_statement_scale_type(self, scale_type):
FILL_FUNC_BODY def test_rotation(self):
rotation = bbob._R(5, sys.maxsize, 'A', b'XAFDAF', 3)
rot... | problem = bbob.DefaultBBOBProblemStatement(10, scale_type=scale_type)
for p in problem.search_space.parameters:
self.assertEqual(p.scale_type, scale_type)
| problem = bbob.DefaultBBOBProblemStatement(10, scale_type=scale_type)
for p in problem.search_space.parameters:
self.assertEqual(p.scale_type, scale_type)
| 10 | 209 | 61 | 270 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,035 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/synthetic/branin.py | 26,27 | 1,70 | _branin | 38,51 | 27,51 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... | a = 1
b = 5.1 / (4 * np.pi**2)
c = 5 / np.pi
r = 6
s = 10
t = 1 / (8 * np.pi)
x1 = x[..., 0]
x2 = x[..., 1]
y = a * (x2 - b * x1**2 + c * x1 - r) ** 2 + s * (1 - t) * np.cos(x1) + s
return y
| """Branin function.
This function can accept batch shapes, although it is typed to return floats
to conform to NumpyExperimenter API.
Args:
x: Shape (B*, 2) array.
Returns:
Shape (B*) array.
"""
a = 1
b = 5.1 / (4 * np.pi**2)
c = 5 / np.pi
r = 6
s = 10
t = 1 / (8 * np.pi)
x1 = x[...... | 65 | 601 | 144 | 745 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... | |
47,036 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/synthetic/branin.py | 62,63 | 51,70 | problem_statement | Branin2DExperimenter | 63,70 | 63,70 | class Branin2DExperimenter(experimenter.Experimenter):
"""2D minimization function. See https://www.sfu.ca/~ssurjano/branin.html."""
def __init__(self):
self._impl = numpy_experimenter.NumpyExperimenter(
_branin, self.problem_statement()
)
def evaluate(self, suggestions: Sequence[vz.Trial]) -> N... | class Branin2DExperimenter(experimenter.Experimenter):
"""2D minimization function. See https://www.sfu.ca/~ssurjano/branin.html."""
def __init__(self):
self._impl = numpy_experimenter.NumpyExperimenter(
_branin, self.problem_statement()
)
def evaluate(self, suggestions: Sequence[vz.Trial]) -> N... | problem = vz.ProblemStatement()
problem.search_space.root.add_float_param("x1", -5, 10)
problem.search_space.root.add_float_param("x2", 0, 15)
problem.metric_information.append(
vz.MetricInformation(name="value", goal=vz.ObjectiveMetricGoal.MINIMIZE)
)
return problem
| problem = vz.ProblemStatement()
problem.search_space.root.add_float_param("x1", -5, 10)
problem.search_space.root.add_float_param("x2", 0, 15)
problem.metric_information.append(
vz.MetricInformation(name="value", goal=vz.ObjectiveMetricGoal.MINIMIZE)
)
return problem
| 15 | 159 | 118 | 277 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,037 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/synthetic/branin_test.py | 27,28 | 25,51 | test_branin_impl | Branin2DExperimenterTest | 28,36 | 28,36 | class Branin2DExperimenterTest(absltest.TestCase):
def test_branin_impl(self):
FILL_FUNC_BODY def test_experimenter_argmin(self):
trial = vz.Trial(parameters={'x1': -np.pi, 'x2': 12.275})
branin.Branin2DExperimenter().evaluate([trial])
self.assertAlmostEqual(
trial.final_measurement_or_die.metri... | class Branin2DExperimenterTest(absltest.TestCase):
def test_branin_impl(self):
FILL_FUNC_BODY def test_experimenter_argmin(self):
trial = vz.Trial(parameters={'x1': -np.pi, 'x2': 12.275})
branin.Branin2DExperimenter().evaluate([trial])
self.assertAlmostEqual(
trial.final_measurement_or_die.metri... | np.testing.assert_allclose(
branin._branin(
np.array([[-np.pi, 12.275], [np.pi, 2.275], [9.42478, 2.475]])
),
0.397887,
atol=1e-5,
)
| np.testing.assert_allclose(
branin._branin(
np.array([[-np.pi, 12.275], [np.pi, 2.275], [9.42478, 2.475]])
),
0.397887,
atol=1e-5,
)
| 12 | 216 | 105 | 321 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,038 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/synthetic/hartmann.py | 27,30 | 1,101 | _hartmann | 30,34 | 30,34 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... | # pylint:disable=invalid-name
return -alpha @ np.exp(-np.sum(A * (x - P) ** 2, axis=1))
| # pylint:disable=invalid-name
return -alpha @ np.exp(-np.sum(A * (x - P) ** 2, axis=1))
| 13 | 1,227 | 47 | 1,274 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... | |
47,039 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/synthetic/hartmann.py | 49,50 | 34,59 | problem_statement | HartmannExperimenter | 50,59 | 50,59 | class HartmannExperimenter(experimenter.Experimenter):
"""General Hartmann minimization function."""
# pylint:disable=invalid-name
def __init__(self, alpha: np.ndarray, A: np.ndarray, P: np.ndarray):
self._dimension = A.shape[-1]
self._impl = numpy_experimenter.NumpyExperimenter(
functools.parti... | class HartmannExperimenter(experimenter.Experimenter):
"""General Hartmann minimization function."""
# pylint:disable=invalid-name
def __init__(self, alpha: np.ndarray, A: np.ndarray, P: np.ndarray):
self._dimension = A.shape[-1]
self._impl = numpy_experimenter.NumpyExperimenter(
functools.parti... | problem = vz.ProblemStatement()
for i in range(1, self._dimension + 1):
problem.search_space.root.add_float_param(f"x{i}", 0, 1)
problem.metric_information.append(
vz.MetricInformation(name="value", goal=vz.ObjectiveMetricGoal.MINIMIZE)
)
return problem
| problem = vz.ProblemStatement()
for i in range(1, self._dimension + 1):
problem.search_space.root.add_float_param(f"x{i}", 0, 1)
problem.metric_information.append(
vz.MetricInformation(name="value", goal=vz.ObjectiveMetricGoal.MINIMIZE)
)
return problem
| 19 | 201 | 115 | 316 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,040 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/synthetic/hartmann_test.py | 28,29 | 26,47 | test_experimenter_argmin | Hartmann3DExperimenterTest | 29,41 | 29,41 | class Hartmann3DExperimenterTest(absltest.TestCase):
def test_experimenter_argmin(self):
FILL_FUNC_BODY def test_experimenter(self):
experimenter_testing.assert_evaluates_random_suggestions(
self, hartmann.Hartmann3DExperimenter()
)
| class Hartmann3DExperimenterTest(absltest.TestCase):
def test_experimenter_argmin(self):
FILL_FUNC_BODY def test_experimenter(self):
experimenter_testing.assert_evaluates_random_suggestions(
self, hartmann.Hartmann3DExperimenter()
)
| trial = vz.Trial(
parameters={
f'x{i+1}': x for i, x in enumerate([0.114614, 0.555649, 0.852547])
}
)
hartmann.Hartmann3DExperimenter().evaluate([trial])
self.assertAlmostEqual(
trial.final_measurement_or_die.metrics.get_value('value', np.nan),
-3.86278,
... | trial = vz.Trial(
parameters={
f'x{i+1}': x for i, x in enumerate([0.114614, 0.555649, 0.852547])
}
)
hartmann.Hartmann3DExperimenter().evaluate([trial])
self.assertAlmostEqual(
trial.final_measurement_or_die.metrics.get_value('value', np.nan),
-3.86278,
... | 22 | 92 | 159 | 251 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,041 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/synthetic/hartmann_test.py | 49,50 | 47,71 | test_experimenter_argmin | Hartmann6DExperimenterTest | 50,65 | 50,65 | class Hartmann6DExperimenterTest(absltest.TestCase):
def test_experimenter_argmin(self):
FILL_FUNC_BODY def test_experimenter(self):
experimenter_testing.assert_evaluates_random_suggestions(
self, hartmann.Hartmann6DExperimenter()
)
| class Hartmann6DExperimenterTest(absltest.TestCase):
def test_experimenter_argmin(self):
FILL_FUNC_BODY def test_experimenter(self):
experimenter_testing.assert_evaluates_random_suggestions(
self, hartmann.Hartmann6DExperimenter()
)
| trial = vz.Trial(
parameters={
f'x{i+1}': x
for i, x in enumerate(
[0.20169, 0.150011, 0.476874, 0.275332, 0.311652, 0.6573]
)
}
)
hartmann.Hartmann6DExperimenter().evaluate([trial])
self.assertAlmostEqual(
trial.final_measureme... | trial = vz.Trial(
parameters={
f'x{i+1}': x
for i, x in enumerate(
[0.20169, 0.150011, 0.476874, 0.275332, 0.311652, 0.6573]
)
}
)
hartmann.Hartmann6DExperimenter().evaluate([trial])
self.assertAlmostEqual(
trial.final_measureme... | 27 | 92 | 197 | 289 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,043 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/synthetic/simplekd.py | 207,209 | 88,230 | optimal_objective | SimpleKDExperimenter | 210,230 | 209,230 | @attrs.define
class SimpleKDExperimenter(experimenter.Experimenter):
"""Simple problem with arbitrary number of parameter of each type.
SimpleKD is a carefully constructed function to trap an underexploring
hillclimb algorithm to local optima. It is parameterized by "best_category",
which optimization algorith... | @attrs.define
class SimpleKDExperimenter(experimenter.Experimenter):
"""Simple problem with arbitrary number of parameter of each type.
SimpleKD is a carefully constructed function to trap an underexploring
hillclimb algorithm to local optima. It is parameterized by "best_category",
which optimization algorith... | values = {}
if self.best_category == 'corner':
values['float'] = [-1.0 for _ in range(self.num_float_param)]
values['int'] = [1 for _ in range(self.num_float_param)]
values['discrete'] = [1 for _ in range(self.num_float_param)]
values['categorical'] = ['corner']
if self.best_categor... | """Computes the optimal objective value of a SimpleKD problem."""
values = {}
if self.best_category == 'corner':
values['float'] = [-1.0 for _ in range(self.num_float_param)]
values['int'] = [1 for _ in range(self.num_float_param)]
values['discrete'] = [1 for _ in range(self.num_float_para... | 89 | 1,469 | 320 | 1,789 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,044 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/experimenters/synthetic/simplekd_test.py | 86,94 | 27,122 | test_optimal_relative_error | SimpleKDExperimenterTest | 94,122 | 94,122 | class SimpleKDExperimenterTest(parameterized.TestCase):
@parameterized.product(
best_category=['corner', 'center', 'mixed'],
output_relative_error=[True, False],
)
def test_sweep(
self,
best_category: simplekd.SimpleKDCategory,
output_relative_error: bool,
) -> None:
experimen... | class SimpleKDExperimenterTest(parameterized.TestCase):
@parameterized.product(
best_category=['corner', 'center', 'mixed'],
output_relative_error=[True, False],
)
def test_sweep(
self,
best_category: simplekd.SimpleKDCategory,
output_relative_error: bool,
) -> None:
experimen... | exptr_simple4d = simplekd.SimpleKDExperimenter(
best_category, output_relative_error=True
)
categorical_values = ['corner', 'center', 'mixed']
discrete_values = (1, 2, 5, 6, 8)
integer_values = [1, 2, 3]
continuous_values = np.linspace(-1, 1, 1001)
opt_rel_error = np.nan
for cat... | exptr_simple4d = simplekd.SimpleKDExperimenter(
best_category, output_relative_error=True
)
categorical_values = ['corner', 'center', 'mixed']
discrete_values = (1, 2, 5, 6, 8)
integer_values = [1, 2, 3]
continuous_values = np.linspace(-1, 1, 1001)
opt_rel_error = np.nan
for cat... | 78 | 706 | 359 | 1,065 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,046 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/benchmarks/runners/benchmark_runner_test.py | 45,75 | 43,194 | test_benchmark_run | BaseRunnerTest | 75,88 | 75,88 | class BaseRunnerTest(parameterized.TestCase):
@parameterized.parameters(
{
'runner': benchmark_runner.BenchmarkRunner(
benchmark_subroutines=[
benchmark_runner.GenerateSuggestions(),
benchmark_runner.EvaluateActiveTrials(),
],
... | class BaseRunnerTest(parameterized.TestCase):
@parameterized.parameters(
{
'runner': benchmark_runner.BenchmarkRunner(
benchmark_subroutines=[
benchmark_runner.GenerateSuggestions(),
benchmark_runner.EvaluateActiveTrials(),
],
... | benchmark_state_factory = _get_designer_benchmark_state_factory()
bench_state = benchmark_state_factory(seed=5)
runner.run(bench_state)
self.assertEmpty(
bench_state.algorithm.supporter.GetTrials(
status_matches=vz.TrialStatus.ACTIVE
)
)
all_trials = bench_state.algor... | benchmark_state_factory = _get_designer_benchmark_state_factory()
bench_state = benchmark_state_factory(seed=5)
runner.run(bench_state)
self.assertEmpty(
bench_state.algorithm.supporter.GetTrials(
status_matches=vz.TrialStatus.ACTIVE
)
)
all_trials = bench_state.algor... | 23 | 1,475 | 174 | 1,649 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
47,047 | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier | /home/hieuvd/dung_repo/raw/stars_03/google_vizier/vizier/vizier/_src/jax/xla_pareto_test.py | 43,44 | 26,62 | test_get_pareto_rank | XlaParetoTest | 44,49 | 44,49 | class XlaParetoTest(absltest.TestCase):
def test_get_pareto(self):
ys = np.random.random([20000, 2])
is_frontier = xla_pareto.is_frontier(ys, num_shards=5)
frontier = xla_pareto.get_frontier(ys, num_shards=5)
# Two implementations should be consistent.
self.assertEqual(frontier.shape[0], np.sum(... | class XlaParetoTest(absltest.TestCase):
def test_get_pareto(self):
ys = np.random.random([20000, 2])
is_frontier = xla_pareto.is_frontier(ys, num_shards=5)
frontier = xla_pareto.get_frontier(ys, num_shards=5)
# Two implementations should be consistent.
self.assertEqual(frontier.shape[0], np.sum(... | ys = np.array([[6, 6], [4, 8], [2, 6], [3, 6], [1, 8], [7, 0], [4, 4],
[9, 6], [5, 10], [8, 10]])
np.testing.assert_array_equal(
xla_pareto.pareto_rank(ys), np.array([2, 2, 6, 5, 3, 2, 5, 0, 1, 0]))
| ys = np.array([[6, 6], [4, 8], [2, 6], [3, 6], [1, 8], [7, 0], [4, 4],
[9, 6], [5, 10], [8, 10]])
np.testing.assert_array_equal(
xla_pareto.pareto_rank(ys), np.array([2, 2, 6, 5, 3, 2, 5, 0, 1, 0]))
| 34 | 399 | 140 | 539 | # Copyright 2024 Google LLC.
#
# 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, ... | # Copyright 2024 Google LLC.
#
# 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, ... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.