Open jason-dai opened 4 years ago
Not the same, their outputs' number of tensors is different. And Census's WND use a Sequential model because this issue https://github.com/intel-analytics/analytics-zoo-internal/issues/1080. @jason-dai
What's the difference between getDeepTensors https://github.com/intel-analytics/analytics-zoo/blob/master/zoo/src/main/scala/com/intel/analytics/zoo/models/recommendation/Utils.scala#L191 and getDeepTensor https://github.com/intel-analytics/analytics-zoo/blob/master/zoo/src/main/scala/com/intel/analytics/zoo/models/recommendation/Utils.scala#L244 Much duplicate in code here @jason-dai @qiuxin2012
e.... getDeepTensors return an array of tensors for indicator, embed, continuous while getDeepTensor returns a combinations of them as a concatenated single tensor... Why do we need both of them? Even if we do need both of them, most of the code can be shared. Weird... @jason-dai @qiuxin2012
row2SampleSequential (https://github.com/intel-analytics/analytics-zoo/blob/master/zoo/src/main/scala/com/intel/analytics/zoo/models/recommendation/Utils.scala#L136) and row2Sample (https://github.com/intel-analytics/analytics-zoo/blob/master/zoo/src/main/scala/com/intel/analytics/zoo/models/recommendation/Utils.scala#L108)
Aren't they the same? We should combine them to one
row2SampleWnD
method