-
Notifications
You must be signed in to change notification settings - Fork 7.8k
/
evaluator.h
109 lines (88 loc) · 3.45 KB
/
evaluator.h
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
// Copyright 2024 The Google Research Authors.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
#ifndef AUTOML_ZERO_EVALUATOR_H_
#define AUTOML_ZERO_EVALUATOR_H_
#include <cstdio>
#include <memory>
#include <random>
#include <vector>
#include "algorithm.h"
#include "task.h"
#include "task.pb.h"
#include "definitions.h"
#include "experiment.pb.h"
#include "fec_cache.h"
#include "random_generator.h"
#include "train_budget.h"
namespace automl_zero {
class Algorithm;
// See base class.
class Evaluator {
public:
Evaluator(
const FitnessCombinationMode fitness_combination_mode,
// Tasks to use. Will be filtered to only keep tasks targeted
// to this worker.
const TaskCollection& task_collection,
// The random generator seed to use for any random operations that
// may be executed by the component function (e.g. VectorRandomInit).
RandomGenerator* rand_gen,
// An cache to avoid reevaluating models that are functionally
// identical. Can be nullptr.
FECCache* functional_cache,
// A train budget to use.
TrainBudget* train_budget,
// Errors larger than this trigger early stopping, as they signal
// models that likely have runnaway behavior.
double max_abs_error);
// If false, suppresses all logging output. Finer grain control
// available through logging flags.
Evaluator(const Evaluator& other) = delete;
Evaluator& operator=(const Evaluator& other) = delete;
// Evaluates a Algorithm by executing it on the tasks. Returns the mean
// fitness.
double Evaluate(const Algorithm& algorithm);
// Get the number of train steps this evaluator has performed.
IntegerT GetNumTrainStepsCompleted() const;
private:
double Execute(const TaskInterface& task, IntegerT num_train_examples,
const Algorithm& algorithm);
template <FeatureIndexT F>
double ExecuteImpl(const Task<F>& task, IntegerT num_train_examples,
const Algorithm& algorithm);
double CapFitness(double fitness);
const FitnessCombinationMode fitness_combination_mode_;
// Contains only task specifications targeted to his worker.
const TaskCollection task_collection_;
TrainBudget* train_budget_;
RandomGenerator* rand_gen_;
std::vector<std::unique_ptr<TaskInterface>> tasks_;
FECCache* functional_cache_;
std::unique_ptr<std::mt19937> functional_cache_bit_gen_owned_;
std::unique_ptr<RandomGenerator> functional_cache_rand_gen_owned_;
RandomGenerator* functional_cache_rand_gen_;
const std::vector<RandomSeedT> first_param_seeds_;
const std::vector<RandomSeedT> first_data_seeds_;
double best_fitness_;
std::shared_ptr<Algorithm> best_algorithm_;
const double max_abs_error_;
IntegerT num_train_steps_completed_;
};
namespace internal {
double CombineFitnesses(
const std::vector<double>& task_fitnesses,
const FitnessCombinationMode mode);
} // namespace internal
} // namespace automl_zero
#endif // AUTOML_ZERO_EVALUATOR_H_