search

zalandoresearch / pytorch-ts

PyTorch based Probabilistic Time Series forecasting framework based on GluonTS backend
MIT License
1.21k stars 191 forks source link

Hyperparameter settings (TemporalFlow) #141

Open hanlaoshi opened 1 year ago

hanlaoshi commented 1 year ago

Thank you for visiting my issue! I have recently been trying to replicate your paper's results regarding using the Solar model on the Traffic dataset. However, I have been encountering some difficulties reproducing the reported results with the hyperparameters provided in the paper.

I was wondering if you could kindly provide me with the hyperparameters used for the Traffic dataset, or if you could advise me on which specific hyperparameters I should focus on adjusting in order to achieve the reported results.

xinyaofan commented 1 year ago

@hanlaoshi i use the electricity dataset and get the results reported in the paper either. ..Have you tried the "electricity-nips" dataset? Could you get the 0.0206+-0.0001 results? I actually crate a new issue just now.