Open zengkaiwei opened 5 years ago
Usually rerank is used in one dataset,but in your rerank.py you rerank source feature with target feature? Why? Or my understanding is fault?
@zengkaiwei We are doing it under in the context of domain adaptation. So we think both datasets matters.
Usually rerank is used in one dataset,but in your rerank.py you rerank source feature with target feature? Why? Or my understanding is fault?