#
# Copyright 2016 The BigDL 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 exp'
# ress or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#
from .base_automodel import BaseAutomodel
[docs]class AutoSeq2Seq(BaseAutomodel):
def __init__(self,
input_feature_num,
output_target_num,
past_seq_len,
future_seq_len,
optimizer,
loss,
metric,
metric_mode=None,
lr=0.001,
lstm_hidden_dim=128,
lstm_layer_num=2,
dropout=0.25,
teacher_forcing=False,
backend="torch",
logs_dir="/tmp/auto_seq2seq",
cpus_per_trial=1,
name="auto_seq2seq",
remote_dir=None,
):
"""
Create an AutoSeq2Seq.
:param input_feature_num: Int. The number of features in the input
:param output_target_num: Int. The number of targets in the output
:param past_seq_len: Int. The number of historical steps used for forecasting.
:param future_seq_len: Int. The number of future steps to forecast.
:param optimizer: String or pyTorch optimizer creator function or
tf.keras optimizer instance.
:param loss: String or pytorch/tf.keras loss instance or pytorch loss creator function.
:param metric: String or customized evaluation metric function.
If string, metric is the evaluation metric name to optimize, e.g. "mse".
If callable function, it signature should be func(y_true, y_pred), where y_true and
y_pred are numpy ndarray. The function should return a float value
as evaluation result.
:param metric_mode: One of ["min", "max"]. "max" means greater metric value is better.
You have to specify metric_mode if you use a customized metric function.
You don't have to specify metric_mode if you use the built-in metric in
bigdl.orca.automl.metrics.Evaluator.
:param lr: float or hp sampling function from a float space. Learning rate.
e.g. hp.choice([0.001, 0.003, 0.01])
:param lstm_hidden_dim: LSTM hidden channel for decoder and encoder.
hp.grid_search([32, 64, 128])
:param lstm_layer_num: LSTM layer number for decoder and encoder.
e.g. hp.randint(1, 4)
:param dropout: float or hp sampling function from a float space. Learning rate. Dropout
rate. e.g. hp.uniform(0.1, 0.3)
:param teacher_forcing: If use teacher forcing in training. e.g. hp.choice([True, False])
:param backend: The backend of the Seq2Seq model. support "keras" and "torch".
:param logs_dir: Local directory to save logs and results. It defaults to
"/tmp/auto_seq2seq"
:param cpus_per_trial: Int. Number of cpus for each trial. It defaults to 1.
:param name: name of the AutoSeq2Seq. It defaults to "auto_seq2seq"
:param remote_dir: String. Remote directory to sync training results and checkpoints. It
defaults to None and doesn't take effects while running in local. While running in
cluster, it defaults to "hdfs:///tmp/{name}".
"""
# todo: support search for past_seq_len.
self.search_space = dict(
input_feature_num=input_feature_num,
output_feature_num=output_target_num,
past_seq_len=past_seq_len,
future_seq_len=future_seq_len,
lstm_hidden_dim=lstm_hidden_dim,
lstm_layer_num=lstm_layer_num,
lr=lr,
dropout=dropout,
teacher_forcing=teacher_forcing
)
self.metric = metric
self.metric_mode = metric_mode
self.backend = backend
self.optimizer = optimizer
self.loss = loss
self._auto_est_config = dict(logs_dir=logs_dir,
resources_per_trial={"cpu": cpus_per_trial},
remote_dir=remote_dir,
name=name)
if self.backend.startswith("torch"):
from bigdl.chronos.model.Seq2Seq_pytorch import model_creator
elif self.backend.startswith("keras"):
from bigdl.chronos.model.tf2.Seq2Seq_keras import model_creator_auto as model_creator
else:
from bigdl.nano.utils.common import invalidInputError
invalidInputError(False,
f"We only support keras and torch as backend,"
f" but got {self.backend}")
self._model_creator = model_creator
super().__init__()