Source code for gluonts.model.simple_feedforward._network

# Copyright 2018, Inc. or its affiliates. All Rights Reserved.
# Licensed under the Apache License, Version 2.0 (the "License").
# You may not use this file except in compliance with the License.
# A copy of the License is located at
# or in the "license" file accompanying this file. This file is distributed
# express or implied. See the License for the specific language governing
# permissions and limitations under the License.

# Standard library imports
from typing import List

# Third-party imports
import mxnet as mx

# First-party imports
from gluonts.block.scaler import MeanScaler, NOPScaler
from gluonts.core.component import validated
from gluonts.distribution import Distribution, DistributionOutput
from gluonts.model.common import Tensor

[docs]class SimpleFeedForwardNetworkBase(mx.gluon.HybridBlock): """ Abstract base class to implement feed-forward networks for probabilistic time series prediction. This class does not implement hybrid_forward: this is delegated to the two subclasses SimpleFeedForwardTrainingNetwork and SimpleFeedForwardPredictionNetwork, that define respectively how to compute the loss and how to generate predictions. Parameters ---------- num_hidden_dimensions Number of hidden nodes in each layer. prediction_length Number of time units to predict. context_length Number of time units that condition the predictions. batch_normalization Whether to use batch normalization. mean_scaling Scale the network input by the data mean and the network output by its inverse. distr_output Distribution to fit. kwargs """ # Needs the validated decorator so that arguments types are checked and # the block can be serialized. @validated() def __init__( self, num_hidden_dimensions: List[int], prediction_length: int, context_length: int, batch_normalization: bool, mean_scaling: bool, distr_output: DistributionOutput, **kwargs, ) -> None: super().__init__(**kwargs) self.num_hidden_dimensions = num_hidden_dimensions self.prediction_length = prediction_length self.context_length = context_length self.batch_normalization = batch_normalization self.mean_scaling = mean_scaling self.distr_output = distr_output with self.name_scope(): self.distr_args_proj = self.distr_output.get_args_proj() self.mlp = mx.gluon.nn.HybridSequential() dims = self.num_hidden_dimensions for layer_no, units in enumerate(dims[:-1]): self.mlp.add(mx.gluon.nn.Dense(units=units, activation="relu")) if self.batch_normalization: self.mlp.add(mx.gluon.nn.BatchNorm()) self.mlp.add(mx.gluon.nn.Dense(units=prediction_length * dims[-1])) self.mlp.add( mx.gluon.nn.HybridLambda( lambda F, o: F.reshape( o, (-1, prediction_length, dims[-1]) ) ) ) self.scaler = MeanScaler() if mean_scaling else NOPScaler()
[docs] def get_distr(self, F, past_target: Tensor) -> Distribution: """ Given past target values, applies the feed-forward network and maps the output to a probability distribution for future observations. Parameters ---------- F past_target Tensor containing past target observations. Shape: (batch_size, context_length, target_dim). Returns ------- Distribution The predicted probability distribution for future observations. """ # (batch_size, seq_len, target_dim) and (batch_size, seq_len, target_dim) scaled_target, target_scale = self.scaler( past_target, F.ones_like(past_target), # TODO: pass the actual observed here ) mlp_outputs = self.mlp(scaled_target) distr_args = self.distr_args_proj(mlp_outputs) return self.distr_output.distribution( distr_args, scale=target_scale.expand_dims(axis=1) )
[docs]class SimpleFeedForwardTrainingNetwork(SimpleFeedForwardNetworkBase): # noinspection PyMethodOverriding,PyPep8Naming
[docs] def hybrid_forward( self, F, past_target: Tensor, future_target: Tensor ) -> Tensor: """ Computes a probability distribution for future data given the past, and returns the loss associated with the actual future observations. Parameters ---------- F past_target Tensor with past observations. Shape: (batch_size, context_length, target_dim). future_target Tensor with future observations. Shape: (batch_size, prediction_length, target_dim). Returns ------- Tensor Loss tensor. Shape: (batch_size, ). """ distr = self.get_distr(F, past_target) # (batch_size, prediction_length, target_dim) loss = distr.loss(future_target) # (batch_size, ) return loss.mean(axis=1)
[docs]class SimpleFeedForwardPredictionNetwork(SimpleFeedForwardNetworkBase): @validated() def __init__( self, num_parallel_samples: int = 100, *args, **kwargs ) -> None: super().__init__(*args, **kwargs) self.num_parallel_samples = num_parallel_samples # noinspection PyMethodOverriding,PyPep8Naming
[docs] def hybrid_forward(self, F, past_target: Tensor) -> Tensor: """ Computes a probability distribution for future data given the past, and draws samples from it. Parameters ---------- F past_target Tensor with past observations. Shape: (batch_size, context_length, target_dim). Returns ------- Tensor Prediction sample. Shape: (samples, batch_size, prediction_length). """ distr = self.get_distr(F, past_target) # (num_samples, batch_size, prediction_length) samples = distr.sample(self.num_parallel_samples) # (batch_size, num_samples, prediction_length) return samples.swapaxes(0, 1)