Skip to content

Freeze thaw sampler

neps.optimizers.bayesian_optimization.acquisition_samplers.freeze_thaw_sampler #

FreezeThawSampler #

FreezeThawSampler(**kwargs)

Bases: AcquisitionSampler

Source code in neps/optimizers/bayesian_optimization/acquisition_samplers/freeze_thaw_sampler.py 
19
20
21
22
23
24
25
26
27
28
def __init__(self, **kwargs):
    super().__init__(**kwargs)
    self.observations = None
    self.b_step = None
    self.n = None
    self.pipeline_space = None
    # args to manage tabular spaces/grid
    self.is_tabular = False
    self.sample_full_table = None
    self.set_sample_full_tabular(True)  # sets flag that samples full table

sample #

sample(
    acquisition_function=None,
    n: int = None,
    set_new_sample_fidelity: int | float = None,
) -> list()

Samples a new set and returns the total set of observed + new configs.

Source code in neps/optimizers/bayesian_optimization/acquisition_samplers/freeze_thaw_sampler.py 
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
def sample(
    self,
    acquisition_function=None,
    n: int = None,
    set_new_sample_fidelity: int | float = None,
) -> list():
    """Samples a new set and returns the total set of observed + new configs."""
    partial_configs = self.observations.get_partial_configs_at_max_seen()
    new_configs = self._sample_new(
        index_from=self.observations.next_config_id(), n=n, ignore_fidelity=False
    )

    def __sample_single_new_tabular(index: int):
        """
        A function to use in a list comprehension to slightly speed up
        the sampling process when self.SAMPLE_TO_DRAW is large
        """
        config = self.pipeline_space.sample(
            patience=self.patience, user_priors=False, ignore_fidelity=False
        )
        config["id"].value = _new_configs[index]
        config.fidelity.value = set_new_sample_fidelity
        return config

    if self.is_tabular:
        _n = n if n is not None else self.SAMPLES_TO_DRAW
        _partial_ids = {conf["id"].value for conf in partial_configs}
        _all_ids = set(self.pipeline_space.custom_grid_table.index.values)

        # accounting for unseen configs only, samples remaining table if flag is set
        max_n = len(_all_ids) + 1 if self.sample_full_table else _n
        _n = min(max_n, len(_all_ids - _partial_ids))

        _new_configs = np.random.choice(
            list(_all_ids - _partial_ids), size=_n, replace=False
        )
        new_configs = [__sample_single_new_tabular(i) for i in range(_n)]
        new_configs = pd.Series(
            new_configs,
            index=np.arange(
                len(partial_configs), len(partial_configs) + len(new_configs)
            ),
        )

    elif set_new_sample_fidelity is not None:
        for config in new_configs:
            config.fidelity.value = set_new_sample_fidelity

    # Deep copy configs for fidelity updates
    partial_configs_list = []
    index_list = []
    for idx, config in partial_configs.items():
        _config = deepcopy(config)
        partial_configs_list.append(_config)
        index_list.append(idx)

    # We build a new series of partial configs to avoid
    # incrementing fidelities multiple times due to pass-by-reference
    partial_configs = pd.Series(partial_configs_list, index=index_list)

    configs = pd.concat([partial_configs, new_configs])

    return configs