AutoTFT Error with Exogenous Features

What happened + What you expected to happen

When training TFT model with exogenous features, an error occurs (below). This error does not occur if exogenous variables are 'None' (eg., 'hist_exog_list'=None). Other models (NHITS, LSTM, MLP, TCN, RNN) do not have this issue.

File "/neuralforecast/neuralforecast/common/_base_model.py", line 55, in _get_temporal_exogenous_cols set(temporal_cols.tolist()) & set(self.hist_exog_list + self.futr_exog_list) TypeError: can only concatenate tuple (not "list") to tuple

Code is provided to replicate the error.

Versions / Dependencies

neuralforecast==1.7.0 (all recent updates pulled)

OS: Springdale Open Enterprise Linux (--> based on Red Hat Enterprise Linux) Browser: N/A Version: 8.7 Moderna

Reproduction script

import torch
import pandas as pd
from datasetsforecast.long_horizon import LongHorizon

from ray import tune
from ray.tune.search.hyperopt import HyperOptSearch

from neuralforecast.auto import AutoRNN, AutoLSTM, AutoTCN, AutoTFT
from neuralforecast.core import NeuralForecast
from neuralforecast.losses.pytorch import MAE

import logging
logging.getLogger("pytorch_lightning").setLevel(logging.WARNING)

def main():
    Y_df = pd.read_csv('/longhorizon/datasets/ili/M/df_y.csv)
    Y_df['ds'] = pd.to_datetime(Y_df['ds'])

    # add exogenous variable to test
    Y_df['exog1'] = 3

    n_time = len(Y_df.ds.unique())
    val_size = int(.2 * n_time)
    test_size = int(.2 * n_time)
    horizon = 96

    # compare tft to lstm baseline
    lstm_config = {
        "hist_exog_list":['exog1'],
        "learning_rate": tune.choice([1e-3]),
        "max_steps": tune.choice([4]),   
        "val_check_steps": tune.choice([2]),
        "input_size": tune.choice([2 * horizon]),
        "encoder_n_layers":tune.choice([2]),
        "encoder_hidden_size":tune.choice([128]),
        "context_size":tune.choice([10]),
        "decoder_hidden_size":tune.choice([128]),
        "decoder_layers":tune.choice([2]),
        "random_seed": tune.choice([1, 2, 3, 4, 5]),
    }

    tft_config = {
        "hist_exog_list":['exog1'],
        "learning_rate": tune.choice([1e-3]),
        "max_steps": tune.choice([4]),   
        "val_check_steps": tune.choice([2]),
        "input_size": tune.choice([2 * horizon]),                                 
        'hidden_size': tune.choice([256, 512]),
        'dropout': tune.choice([0]),
        'attn_dropout': tune.choice([0]),
        "random_seed": tune.choice([1, 2, 3, 4, 5]),
    }

    models = [AutoLSTM(h=horizon,
                       loss=MAE(),
                           config=lstm_config,
                           search_alg=HyperOptSearch(),
                           num_samples=1),
              AutoTFT(h=horizon,
                      loss=MAE(),
                      config=tft_config,
                      search_alg=HyperOptSearch(),
                      num_samples=1),
             ]

    fcst = NeuralForecast(models=models, freq='7D') #freq='15min'

    fcst_df = fcst.cross_validation(df=Y_df, val_size=val_size,test_size=test_size, n_windows=None)

if __name__=='__main__':
    print('testing')
    main()

Issue Severity

High: It blocks me from completing my task.

Nixtla / neuralforecast