Vanilla Transformer

Vanilla Transformer, following implementation of the Informer paper, used as baseline. The architecture has three distinctive features:

Full-attention mechanism with O(L^2) time and memory complexity.
Classic encoder-decoder proposed by Vaswani et al. (2017) with a multi-head attention mechanism.
An MLP multi-step decoder that predicts long time-series sequences in a single forward operation rather than step-by-step.

The Vanilla Transformer model utilizes a three-component approach to define its embedding:

It employs encoded autoregressive features obtained from a convolution network.
It uses window-relative positional embeddings derived from harmonic functions.
Absolute positional embeddings obtained from calendar features are utilized.

References

Haoyi Zhou, Shanghang Zhang, Jieqi Peng, Shuai Zhang, Jianxin Li, Hui Xiong, Wancai Zhang. “Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting”

Figure 1. Transformer Architecture.

Vanilla Transformer

`VanillaTransformer`

VanillaTransformer(
    h,
    input_size,
    stat_exog_list=None,
    hist_exog_list=None,
    futr_exog_list=None,
    cat_exog_list=None,
    categorical_cardinalities=None,
    cat_emb_dim="fastai",
    exclude_insample_y=False,
    decoder_input_size_multiplier=0.5,
    hidden_size=128,
    dropout=0.05,
    n_head=4,
    conv_hidden_size=32,
    activation="gelu",
    encoder_layers=2,
    decoder_layers=1,
    loss=MAE(),
    valid_loss=None,
    max_steps=5000,
    learning_rate=0.0001,
    num_lr_decays=-1,
    early_stop_patience_steps=-1,
    val_monitor="ptl/val_loss",
    val_check_steps=100,
    batch_size=32,
    valid_batch_size=None,
    windows_batch_size=1024,
    inference_windows_batch_size=1024,
    start_padding_enabled=False,
    training_data_availability_threshold=0.0,
    step_size=1,
    scaler_type="identity",
    random_seed=1,
    drop_last_loader=False,
    alias=None,
    optimizer=None,
    optimizer_kwargs=None,
    lr_scheduler=None,
    lr_scheduler_kwargs=None,
    dataloader_kwargs=None,
    **trainer_kwargs
)

Bases: BaseModel VanillaTransformer Vanilla Transformer, following implementation of the Informer paper, used as baseline. The architecture has three distinctive features:

Full-attention mechanism with O(L^2) time and memory complexity.
An MLP multi-step decoder that predicts long time-series sequences in a single forward operation rather than step-by-step.

The Vanilla Transformer model utilizes a three-component approach to define its embedding:

It employs encoded autoregressive features obtained from a convolution network.
It uses window-relative positional embeddings derived from harmonic functions.
Absolute positional embeddings obtained from calendar features are utilized.

Parameters:

Name	Type	Description	Default
`h`	`int`	forecast horizon.	required
`input_size`	`int`	maximum sequence length for truncated train backpropagation.	required
`stat_exog_list`	`str list`	static exogenous columns.	`None`
`hist_exog_list`	`str list`	historic exogenous columns.	`None`
`futr_exog_list`	`str list`	future exogenous columns.	`None`
`cat_exog_list`	`str list`	exogenous columns (from `hist_exog_list` / `futr_exog_list`) to embed instead of scale.	`None`
`categorical_cardinalities`	`dict`	mapping from each categorical column to its number of distinct categories.	`None`
`cat_emb_dim`	`str or int`	categorical embedding size strategy (‘fastai’, ‘sqrt’, ‘half’) or an explicit integer.	`‘fastai’`
`exclude_insample_y`	`bool`	whether to exclude the target variable from the input.	`False`
`decoder_input_size_multiplier`	`float`	.	`0.5`
`hidden_size`	`int`	units of embeddings and encoders.	`128`
`dropout`	`float`	dropout throughout Informer architecture.	`0.05`
`n_head`	`int`	controls number of multi-head’s attention.	`4`
`conv_hidden_size`	`int`	channels of the convolutional encoder.	`32`
`activation`	`str`	activation from [‘ReLU’, ‘Softplus’, ‘Tanh’, ‘SELU’, ‘LeakyReLU’, ‘PReLU’, ‘Sigmoid’, ‘GELU’].	`‘gelu’`
`encoder_layers`	`int`	number of layers for the TCN encoder.	`2`
`decoder_layers`	`int`	number of layers for the MLP decoder.	`1`
`loss`	`PyTorch module`	instantiated train loss class from losses collection.	`MAE()`
`valid_loss`	`PyTorch module`	instantiated valid loss class from losses collection.	`None`
`max_steps`	`int`	maximum number of training steps.	`5000`
`learning_rate`	`float`	Learning rate between (0, 1).	`0.0001`
`num_lr_decays`	`int`	Number of learning rate decays, evenly distributed across max_steps.	`-1`
`early_stop_patience_steps`	`int`	Number of validation iterations before early stopping.	`-1`
`val_monitor`	`str`	metric to monitor for early stopping. Valid options: “ptl/val_loss”, “valid_loss”, “train_loss”. Default: “ptl/val_loss”.	`‘ptl/val_loss’`
`val_check_steps`	`int`	Number of training steps between every validation loss check.	`100`
`batch_size`	`int`	number of different series in each batch.	`32`
`valid_batch_size`	`int`	number of different series in each validation and test batch, if None uses batch_size.	`None`
`windows_batch_size`	`int`	number of windows to sample in each training batch, default uses all.	`1024`
`inference_windows_batch_size`	`int`	number of windows to sample in each inference batch.	`1024`
`start_padding_enabled`	`bool`	if True, the model will pad the time series with zeros at the beginning, by input size.	`False`
`training_data_availability_threshold`	`Union[float, List[float]]`	minimum fraction of valid data points required for training windows. Single float applies to both insample and outsample; list of two floats specifies [insample_fraction, outsample_fraction]. Default 0.0 allows windows with only 1 valid data point (current behavior).	`0.0`
`step_size`	`int`	step size between each window of temporal data.	`1`
`scaler_type`	`str`	type of scaler for temporal inputs normalization see temporal scalers.	`‘identity’`
`random_seed`	`int`	random_seed for pytorch initializer and numpy generators.	`1`
`drop_last_loader`	`bool`	if True `TimeSeriesDataLoader` drops last non-full batch.	`False`
`alias`	`str`	optional, Custom name of the model.	`None`
`optimizer`	`Subclass of ‘torch.optim.Optimizer’`	optional, user specified optimizer instead of the default choice (Adam).	`None`
`optimizer_kwargs`	`dict`	optional, list of parameters used by the user specified `optimizer`.	`None`
`lr_scheduler`	`Subclass of ‘torch.optim.lr_scheduler.LRScheduler’`	optional, user specified lr_scheduler instead of the default choice (StepLR).	`None`
`lr_scheduler_kwargs`	`dict`	optional, list of parameters used by the user specified `lr_scheduler`.	`None`
`dataloader_kwargs`	`dict`	optional, list of parameters passed into the PyTorch Lightning dataloader by the `TimeSeriesDataLoader`.	`None`
`**trainer_kwargs`	`int`	keyword trainer arguments inherited from PyTorch Lightning’s trainer.

`VanillaTransformer.fit`

fit(
    dataset, val_size=0, test_size=0, random_seed=None, distributed_config=None
)

Fit. The fit method, optimizes the neural network’s weights using the initialization parameters (learning_rate, windows_batch_size, …) and the loss function as defined during the initialization. Within fit we use a PyTorch Lightning Trainer that inherits the initialization’s self.trainer_kwargs, to customize its inputs, see PL’s trainer arguments. The method is designed to be compatible with SKLearn-like classes and in particular to be compatible with the StatsForecast library. By default the model is not saving training checkpoints to protect disk memory, to get them change enable_checkpointing=True in __init__. Parameters:

Name	Type	Description	Default
`dataset`	`TimeSeriesDataset`	NeuralForecast’s `TimeSeriesDataset`, see documentation.	required
`val_size`	`int`	Validation size for temporal cross-validation.	`0`
`random_seed`	`int`	Random seed for pytorch initializer and numpy generators, overwrites model.init’s.	`None`
`test_size`	`int`	Test size for temporal cross-validation.	`0`

Returns:

Type	Description
None

`VanillaTransformer.predict`

predict(
    dataset,
    test_size=None,
    step_size=1,
    random_seed=None,
    quantiles=None,
    h=None,
    explainer_config=None,
    **data_module_kwargs
)

Predict. Neural network prediction with PL’s Trainer execution of predict_step. Parameters:

Name	Type	Description	Default
`dataset`	`TimeSeriesDataset`	NeuralForecast’s `TimeSeriesDataset`, see documentation.	required
`test_size`	`int`	Test size for temporal cross-validation.	`None`
`step_size`	`int`	Step size between each window.	`1`
`random_seed`	`int`	Random seed for pytorch initializer and numpy generators, overwrites model.init’s.	`None`
`quantiles`	`list`	Target quantiles to predict.	`None`
`h`	`int`	Prediction horizon, if None, uses the model’s fitted horizon. Defaults to None.	`None`
`explainer_config`	`dict`	configuration for explanations.	`None`
`**data_module_kwargs`	`dict`	PL’s TimeSeriesDataModule args, see documentation.

Returns:

Type	Description
None

Usage Example

import pandas as pd
import matplotlib.pyplot as plt

from neuralforecast import NeuralForecast
from neuralforecast.models import VanillaTransformer
from neuralforecast.utils import AirPassengersPanel, AirPassengersStatic

Y_train_df = AirPassengersPanel[AirPassengersPanel.ds<AirPassengersPanel['ds'].values[-12]] # 132 train
Y_test_df = AirPassengersPanel[AirPassengersPanel.ds>=AirPassengersPanel['ds'].values[-12]].reset_index(drop=True) # 12 test

model = VanillaTransformer(h=12,
                 input_size=24,
                 hidden_size=16,
                 conv_hidden_size=32,
                 n_head=2,
                 loss=MAE(),
                 scaler_type='robust',
                 learning_rate=1e-3,
                 max_steps=500,
                 val_check_steps=50,
                 early_stop_patience_steps=2)

nf = NeuralForecast(
    models=[model],
    freq='ME'
)
nf.fit(df=Y_train_df, static_df=AirPassengersStatic, val_size=12)
forecasts = nf.predict(futr_df=Y_test_df)

Y_hat_df = forecasts.reset_index(drop=False).drop(columns=['unique_id','ds'])
plot_df = pd.concat([Y_test_df, Y_hat_df], axis=1)
plot_df = pd.concat([Y_train_df, plot_df])

if model.loss.is_distribution_output:
    plot_df = plot_df[plot_df.unique_id=='Airline1'].drop('unique_id', axis=1)
    plt.plot(plot_df['ds'], plot_df['y'], c='black', label='True')
    plt.plot(plot_df['ds'], plot_df['VanillaTransformer-median'], c='blue', label='median')
    plt.fill_between(x=plot_df['ds'][-12:], 
                    y1=plot_df['VanillaTransformer-lo-90'][-12:].values, 
                    y2=plot_df['VanillaTransformer-hi-90'][-12:].values,
                    alpha=0.4, label='level 90')
    plt.grid()
    plt.legend()
    plt.plot()
else:
    plot_df = plot_df[plot_df.unique_id=='Airline1'].drop('unique_id', axis=1)
    plt.plot(plot_df['ds'], plot_df['y'], c='black', label='True')
    plt.plot(plot_df['ds'], plot_df['VanillaTransformer'], c='blue', label='Forecast')
    plt.legend()
    plt.grid()

Getting Started

Capabilities

Tutorials

Use cases

API Reference

Vanilla Transformer

Vanilla Transformer

`VanillaTransformer`

`VanillaTransformer.fit`

`VanillaTransformer.predict`

Usage Example