train_utils.py 7.46 KB
Newer Older
1
2
3
4
5
6
7
8
9
# @author lucasmiranda42
# encoding: utf-8
# module deepof

"""

Simple utility functions used in deepof example scripts. These are not part of the main package

"""
10
from datetime import date, datetime
11

12
from kerastuner import BayesianOptimization, Hyperband
13
from kerastuner import HyperParameters
14
from kerastuner_tensorboard_logger import TensorBoardLogger
15
from typing import Tuple, Union, Any, List
16
17
18
19
20
21
22
import deepof.hypermodels
import deepof.model_utils
import numpy as np
import os
import pickle
import tensorflow as tf

23
24
hp = HyperParameters()

25

26
def load_hparams(hparams):
27
28
29
30
31
32
33
34
    """Loads hyperparameters from a custom dictionary pickled on disc.
    Thought to be used with the output of hyperparameter_tuning.py"""

    if hparams is not None:
        with open(hparams, "rb") as handle:
            hparams = pickle.load(handle)
    else:
        hparams = {
35
36
37
38
39
            "bidirectional_merge": "ave",
            "clipvalue": 1.0,
            "dense_activation": "relu",
            "dense_layers_per_branch": 1,
            "dropout_rate": 1e-3,
40
            "learning_rate": 1e-3,
41
42
43
            "units_conv": 160,
            "units_dense2": 120,
            "units_lstm": 300,
44
45
46
47
48
49
50
51
52
53
54
        }
    return hparams


def load_treatments(train_path):
    """Loads a dictionary containing the treatments per individual,
    to be loaded as metadata in the coordinates class"""
    try:
        with open(
            os.path.join(
                train_path,
55
                [i for i in os.listdir(train_path) if i.endswith(".pkl")][0],
56
57
58
59
60
61
62
63
64
65
66
            ),
            "rb",
        ) as handle:
            treatment_dict = pickle.load(handle)
    except IndexError:
        treatment_dict = None

    return treatment_dict


def get_callbacks(
lucas_miranda's avatar
lucas_miranda committed
67
68
69
70
71
72
    X_train: np.array,
    batch_size: int,
    cp: bool,
    variational: bool,
    predictor: float,
    loss: str,
73
) -> List[Union[Any]]:
74
    """Generates callbacks for model training, including:
75
76
77
78
    - run_ID: run name, with coarse parameter details;
    - tensorboard_callback: for real-time visualization;
    - cp_callback: for checkpoint saving,
    - onecycle: for learning rate scheduling"""
79

80
    run_ID = "{}{}{}_{}".format(
81
82
83
84
85
86
87
        ("GMVAE" if variational else "AE"),
        ("P" if predictor > 0 and variational else ""),
        ("_loss={}".format(loss) if variational else ""),
        datetime.now().strftime("%Y%m%d-%H%M%S"),
    )

    log_dir = os.path.abspath("logs/fit/{}".format(run_ID))
88
    tensorboard_callback = tf.keras.callbacks.TensorBoard(
89
90
91
        log_dir=log_dir,
        histogram_freq=1,
        profile_batch=2,
92
93
94
    )

    onecycle = deepof.model_utils.one_cycle_scheduler(
95
96
        X_train.shape[0] // batch_size * 250,
        max_rate=0.005,
97
98
    )

99
100
101
102
103
104
105
106
107
108
109
110
111
    callbacks = [run_ID, tensorboard_callback, onecycle]

    if cp:
        cp_callback = tf.keras.callbacks.ModelCheckpoint(
            "./logs/checkpoints/" + run_ID + "/cp-{epoch:04d}.ckpt",
            verbose=1,
            save_best_only=False,
            save_weights_only=True,
            save_freq="epoch",
        )
        callbacks.append(cp_callback)

    return callbacks
112
113
114


def tune_search(
115
    data: List[np.array],
116
    encoding_size: int,
117
118
    hypertun_trials: int,
    hpt_type: str,
119
120
    hypermodel: str,
    k: int,
121
    kl_warmup_epochs: int,
122
    loss: str,
123
    mmd_warmup_epochs: int,
124
    overlap_loss: float,
125
    pheno_class: float,
126
127
    predictor: float,
    project_name: str,
128
    callbacks: List,
129
    n_epochs: int = 30,
130
    n_replicas: int = 1,
131
) -> Union[bool, Tuple[Any, Any]]:
132
133
    """Define the search space using keras-tuner and bayesian optimization

134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
    Parameters:
        - train (np.array): dataset to train the model on
        - test (np.array): dataset to validate the model on
        - hypertun_trials (int): number of Bayesian optimization iterations to run
        - hpt_type (str): specify one of Bayesian Optimization (bayopt) and Hyperband (hyperband)
        - hypermodel (str): hypermodel to load. Must be one of S2SAE (plain autoencoder)
        or S2SGMVAE (Gaussian Mixture Variational autoencoder).
        - k (int) number of components of the Gaussian Mixture
        - loss (str): one of [ELBO, MMD, ELBO+MMD]
        - overlap_loss (float): assigns as weight to an extra loss term which
        penalizes overlap between GM components
        - pheno_class (float): adds an extra regularizing neural network to the model,
        which tries to predict the phenotype of the animal from which the sequence comes
        - predictor (float): adds an extra regularizing neural network to the model,
        which tries to predict the next frame from the current one
        - project_name (str): ID of the current run
        - callbacks (list): list of callbacks for the training loop
        - n_epochs (int): optional. Number of epochs to train each run for
        - n_replicas (int): optional. Number of replicas per parameter set. Higher values
         will yield more robust results, but will affect performance severely

    Returns:
        - best_hparams (dict): dictionary with the best retrieved hyperparameters
        - best_run (tf.keras.Model): trained instance of the best model found
158
159
160

    """

161
162
    X_train, y_train, X_val, y_val = data

163
164
165
166
    assert hpt_type in ["bayopt", "hyperband"], (
        "Invalid hyperparameter tuning framework. " "Select one of bayopt and hyperband"
    )

lucas_miranda's avatar
lucas_miranda committed
167
    if hypermodel == "S2SAE":  # pragma: no cover
168
169
170
        assert (
            predictor == 0.0 and pheno_class == 0.0
        ), "Prediction branches are only available for variational models. See documentation for more details"
171
        hypermodel = deepof.hypermodels.SEQ_2_SEQ_AE(input_shape=X_train.shape)
172
173
174

    elif hypermodel == "S2SGMVAE":
        hypermodel = deepof.hypermodels.SEQ_2_SEQ_GMVAE(
175
            input_shape=X_train.shape,
176
            encoding=encoding_size,
177
            kl_warmup_epochs=kl_warmup_epochs,
178
            loss=loss,
179
            mmd_warmup_epochs=mmd_warmup_epochs,
180
            number_of_components=k,
181
            overlap_loss=overlap_loss,
182
            phenotype_predictor=pheno_class,
183
            predictor=predictor,
184
        )
lucas_miranda's avatar
lucas_miranda committed
185

186
187
188
    else:
        return False

189
190
191
192
193
194
195
196
197
198
199
200
201
202
    hpt_params = {
        "hypermodel": hypermodel,
        "executions_per_trial": n_replicas,
        "logger": TensorBoardLogger(metrics=["val_mae"], logdir="./logs/hparams"),
        "objective": "val_mae",
        "project_name": project_name,
        "seed": 42,
        "tune_new_entries": True,
    }

    if hpt_type == "hyperband":
        tuner = Hyperband(
            directory="HyperBandx_{}_{}".format(loss, str(date.today())),
            max_epochs=hypertun_trials,
203
            hyperband_iterations=3,
lucas_miranda's avatar
lucas_miranda committed
204
            factor=2,
205
206
207
208
209
210
211
212
            **hpt_params
        )
    else:
        tuner = BayesianOptimization(
            directory="BayOpt_{}_{}".format(loss, str(date.today())),
            max_trials=hypertun_trials,
            **hpt_params
        )
213
214
215

    print(tuner.search_space_summary())

216
217
218
219
220
221
222
223
224
225
226
    Xs, ys = [X_train], [X_train]
    Xvals, yvals = [X_val], [X_val]

    if predictor > 0.0:
        Xs, ys = X_train[:-1], [X_train[:-1], X_train[1:]]
        Xvals, yvals = X_val[:-1], [X_val[:-1], X_val[1:]]

    if pheno_class > 0.0:
        ys += [y_train]
        yvals += [y_val]

227
    tuner.search(
228
229
        Xs,
        ys,
230
        epochs=n_epochs,
231
        validation_data=(Xvals, yvals),
232
233
        verbose=1,
        batch_size=256,
lucas_miranda's avatar
lucas_miranda committed
234
        callbacks=callbacks,
235
236
237
238
239
    )

    best_hparams = tuner.get_best_hyperparameters(num_trials=1)[0]
    best_run = tuner.hypermodel.build(best_hparams)

lucas_miranda's avatar
lucas_miranda committed
240
241
    print(tuner.results_summary())

lucas_miranda's avatar
lucas_miranda committed
242
    return best_hparams, best_run