[Major] Combined future regressor layer arguments into one single integer list regressors_layers like ar_layers

[kneureither]

🔬 Background

• fixes issue Combine Future regressor NN arguments #1375

🔮 Key changes

Removed the arguments future_regressors_num_hidden_layers and future_regressors_d_hidden and replaced them with a single argument future_regressors_layers : Optional[List[int]]. (following the practice of ar_layers)

• length holds the number of hidden layers
• values hold the dimension of each layer

📋 Review Checklist

• I have performed a self-review of my own code.
• I have commented my code, added docstrings and data types to function definitions.
• I have added pytests to check whether my feature / fix works.

[ourownstory]

Labelled as Major as it is a breaking change.

[ourownstory]

Thank you - Looking good overall!
I left few suggestions.

[ourownstory]

to avoid confusion, phase array of.. as list of ...
Should also state default values.
Looks like we currently have [4,4]. We might want roll this back to no hidden layers by default. Thoughts?

[ourownstory]

The default for this should be identical or lower than ar_layers and lagged_reg_layers

[kneureither]

I simply used the previous default, but agree it makes sense to set the default to []. Done that.

[github-actions]

Model Benchmark

Benchmark	Metric	main	current	diff
AirPassengers	MAE_val	30.8317	30.8317	0.0%	✅
AirPassengers	RMSE_val	31.8177	31.8177	0.0%	✅
AirPassengers	Loss_val	0.01301	0.01301	0.0%	✅
AirPassengers	train_loss	0.00071	0.00071	0.0%	✅
AirPassengers	reg_loss	0	0	0.0%	✅
AirPassengers	MAE	6.88164	6.88164	0.0%	✅
AirPassengers	RMSE	8.92097	8.92097	0.0%	✅
AirPassengers	Loss	0.00076	0.00076	0.0%	✅
AirPassengers	time	17.5903	18.01	2.39%	✅
EnergyPriceDaily	MAE_val	5.64183	5.64183	0.0%	✅
EnergyPriceDaily	RMSE_val	7.19915	7.19915	0.0%	✅
EnergyPriceDaily	Loss_val	0.02893	0.02893	0.0%	✅
EnergyPriceDaily	train_loss	0.02957	0.02957	0.0%	✅
EnergyPriceDaily	reg_loss	0	0	0.0%	✅
EnergyPriceDaily	MAE	6.39657	6.39657	0.0%	✅
EnergyPriceDaily	RMSE	8.56213	8.56213	0.0%	✅
EnergyPriceDaily	Loss	0.02936	0.02936	0.0%	✅
EnergyPriceDaily	time	85.4153	89.36	4.62%	⚠️
YosemiteTemps	MAE_val	0.59751	0.59751	0.0%	✅
YosemiteTemps	RMSE_val	0.88859	0.88859	0.0%	✅
YosemiteTemps	Loss_val	0.00046	0.00046	0.0%	✅
YosemiteTemps	train_loss	0.00126	0.00126	0.0%	✅
YosemiteTemps	reg_loss	0	0	0.0%	✅
YosemiteTemps	MAE	0.97464	0.97464	0.0%	✅
YosemiteTemps	RMSE	1.70813	1.70813	0.0%	✅
YosemiteTemps	Loss	0.00126	0.00126	0.0%	✅
YosemiteTemps	time	334.987	350.09	4.51%	⚠️
PeytonManning	MAE_val	0.35542	0.35542	0.0%	✅
PeytonManning	RMSE_val	0.50403	0.50403	0.0%	✅
PeytonManning	Loss_val	0.01803	0.01803	0.0%	✅
PeytonManning	train_loss	0.01466	0.01466	0.0%	✅
PeytonManning	reg_loss	0	0	0.0%	✅
PeytonManning	MAE	0.34755	0.34755	0.0%	✅
PeytonManning	RMSE	0.49449	0.49449	0.0%	✅
PeytonManning	Loss	0.01465	0.01465	0.0%	✅
PeytonManning	time	100.415	110.75	10.29%	❌

\n

Model training plots

\n ## Model Training ### PeytonManning  ### YosemiteTemps  ### AirPassengers  ### EnergyPriceDaily  \n