search

rakshitha123 / TSForecasting

This repository contains the implementations related to the experiments of a set of publicly available datasets that are used in the time series forecasting research space.
https://forecastingdata.org/
Other
206 stars 44 forks source link

Default trainer from Gluon package in the deep_learning_experiments script #10

Closed BoudabousSafa closed 2 years ago

BoudabousSafa commented 2 years ago

Hello, thank you for making this benchmark available. I noticed that the deep learning implementation relies on the Gluon package and in the train_model function, the trainer is always used with the defaulted parameters. I wonder if it is not better to adjust some params to the considered dataset (basically the num_batches_per_epoch parameter).

rakshitha123 commented 2 years ago

Hello,

Yes, the performance of the models can be improved by tuning their hyperparameters with each dataset. But in our work, we executed all models as benchmarks and thus, we executed all of them with their default parameters. We have mentioned that in our paper as well.

BoudabousSafa commented 2 years ago

Thanks @rakshitha123 for your response.