The explanation of training mask and eval mask

[chenxiaodanhit]

Hi Ivan,

Sorry to bother you. I am confused with the training_mask and eval_mask. May I understand that the training_mask is the mask of input which represents the missing value in input and eval_mask is the mask of target, which stands for the visible ground truth. If I want to conduct forecasting task, is it suitable to only change the parameters horizon and delay in Imputation_dataset.py? Or could you please give some advices for how to build forecasting dataset? Thank you for your help!

[marshka]

Hi,

in the case of forecasting, I would suggest using the SpatioTemporalDataset instead. The ImputationDataset is thought to reconstruct missing values inside a window (the one given to the model). If you want to forecast a complete sequence but then you have missing values in the target label, then my advice is to simply use the mask to compute your error only on valid values.

[chenxiaodanhit]

Thank you for your patience! But I met a new question :

TypeError: init() got multiple values for argument eval_mask in 'Imputationdataset.py`.

Could you please help me how to solve it? Thanks.

[marshka]

Hi, I cannot replicate this issue, can you give me more information about it?

[chenxiaodanhit]

Thank you for your help! I find the problem. It is because the version is not suit. Thank you again!

[PhoebeWenpp]

TypeError: init() got multiple values for argument eval_mask in 'Imputationdataset.py` Hi，I met the same question，Could you please help me how to solve it? Thanks.