martinb-ai
martinb-ai
Gotcha, will update!
Our PR does automatic backtesting but does not fix this bug.the PR assumes the user wants to test the whole series they load in. @colin99d we should probably add in...
That is a good thing to add. We will bring it in for v2!
https://github.com/OpenBB-finance/OpenBBTerminal/pull/1933 Should resolve this issue with the new overhaul of forecasting models and backtesting capabilities.
Yes, working on it! Update- Intro page is done.
It shouldnt affect too much. Will have to do testing for sure !
Absolutely! I split all datasets into train/Val. Thanks for the suggestion. I will supply the entire covariate time series to match with the validation. How does one accomplish early stopping...
This is great @gdevos010 Thank you! I guess I will set up a dictionary of multiple hyperparameters to experiment.
Awesome will do! Thank you @hrzn
Hello @dennisbader I am able to reproduce the bug: ```python model_nbeats = NBEATSModel( input_chunk_length=30, output_chunk_length=7, generic_architecture=True, num_stacks=10, num_blocks=1, num_layers=4, layer_widths=512, n_epochs=100, nr_epochs_val_period=1, batch_size=800, model_name="nbeats_run", force_reset=True, pl_trainer_kwargs={ "accelerator": "gpu", "gpus": [0]...