train_utils.py 7.03 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
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
35
36
37
38
    """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 = {
            "units_conv": 256,
            "units_lstm": 256,
            "units_dense2": 64,
            "dropout_rate": 0.25,
39
            "encoding": 16,
40
41
42
43
44
45
46
47
48
49
50
51
            "learning_rate": 1e-3,
        }
    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,
52
                [i for i in os.listdir(train_path) if i.endswith(".pkl")][0],
53
54
55
56
57
58
59
60
61
62
63
            ),
            "rb",
        ) as handle:
            treatment_dict = pickle.load(handle)
    except IndexError:
        treatment_dict = None

    return treatment_dict


def get_callbacks(
64
    X_train: np.array, batch_size: int, cp: bool, variational: bool, predictor: float, loss: str,
65
) -> List[Union[Any]]:
66
67
68
69
70
71
    """Generates callbacks for model training, including:
        - run_ID: run name, with coarse parameter details;
        - tensorboard_callback: for real-time visualization;
        - cp_callback: for checkpoint saving,
        - onecycle: for learning rate scheduling"""

72
    run_ID = "{}{}{}_{}".format(
73
74
75
76
77
78
79
        ("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))
80
    tensorboard_callback = tf.keras.callbacks.TensorBoard(
81
82
83
84
85
86
87
        log_dir=log_dir, histogram_freq=1, profile_batch=2,
    )

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

88
89
90
91
92
93
94
95
96
97
98
99
100
    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
101
102
103


def tune_search(
104
    data: List[np.array],
105
106
107
    bayopt_trials: int,
    hypermodel: str,
    k: int,
108
    kl_warmup_epochs: int,
109
    loss: str,
110
    mmd_warmup_epochs: int,
111
    overlap_loss: float,
112
    pheno_class: float,
113
114
    predictor: float,
    project_name: str,
115
    callbacks: List,
116
    n_epochs: int = 30,
117
    n_replicas: int = 1,
118
) -> Union[bool, Tuple[Any, Any]]:
119
120
121
122
123
124
    """Define the search space using keras-tuner and bayesian optimization

        Parameters:
            - train (np.array): dataset to train the model on
            - test (np.array): dataset to validate the model on
            - bayopt_trials (int): number of Bayesian optimization iterations to run
125
126
            - hypermodel (str): hypermodel to load. Must be one of S2SAE (plain autoencoder)
            or S2SGMVAE (Gaussian Mixture Variational autoencoder).
127
128
129
130
            - 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
131
132
            - 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
133
134
135
136
            - 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
137
138
139
            - 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
140
141
142
143
144
145
146

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

    """

147
148
    X_train, y_train, X_val, y_val = data

lucas_miranda's avatar
lucas_miranda committed
149
    if hypermodel == "S2SAE":  # pragma: no cover
150
151
152
        assert (
            predictor == 0.0 and pheno_class == 0.0
        ), "Prediction branches are only available for variational models. See documentation for more details"
153
        hypermodel = deepof.hypermodels.SEQ_2_SEQ_AE(input_shape=X_train.shape)
154
155
156

    elif hypermodel == "S2SGMVAE":
        hypermodel = deepof.hypermodels.SEQ_2_SEQ_GMVAE(
157
            input_shape=X_train.shape,
158
            kl_warmup_epochs=kl_warmup_epochs,
159
            loss=loss,
160
            mmd_warmup_epochs=mmd_warmup_epochs,
161
            number_of_components=k,
162
            overlap_loss=overlap_loss,
163
            phenotype_predictor=pheno_class,
164
            predictor=predictor,
165
        )
lucas_miranda's avatar
lucas_miranda committed
166

167
168
169
170
171
    else:
        return False

    tuner = BayesianOptimization(
        hypermodel,
172
        directory="BayesianOptx_{}_{}".format(loss, str(date.today())),
173
        executions_per_trial=n_replicas,
174
175
        logger=TensorBoardLogger(metrics=["val_mae"], logdir="./logs/hparams"),
        max_trials=bayopt_trials,
176
177
        objective="val_mae",
        project_name=project_name,
178
        seed=42,
179
        tune_new_entries=True,
180
181
182
183
    )

    print(tuner.search_space_summary())

184
185
186
187
188
189
190
191
192
193
194
    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]

195
    tuner.search(
196
197
        Xs,
        ys,
198
        epochs=n_epochs,
199
        validation_data=(Xvals, yvals),
200
201
        verbose=1,
        batch_size=256,
lucas_miranda's avatar
lucas_miranda committed
202
        callbacks=callbacks,
203
204
205
206
207
    )

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

lucas_miranda's avatar
lucas_miranda committed
208
209
    print(tuner.results_summary())

210
211
212
213
214
215
216
217
    return best_hparams, best_run


# TODO:
#    - load_treatments should be part of the main data module. If available in the main directory,
#    a table (preferrable in csv) should be loaded as metadata of the coordinates automatically.
#    This becomes particularly important por the supervised models that include phenotype classification
#    alongside the encoding.