Sense-X/X-Temporal

In linear_sampler(data, bias), bias 'shape should N*T*G?

cvnovice95 opened this issue · 1 comments

cvnovice95 commented

In linear_sampler(), bias'shape is NTG, why is it not N*G?

deepcs233 commented

In our setting, bias is different according to its location in the temporal dimension. For detail, you can refer to our paper TIN.