search

easezyc / deep-transfer-learning

A collection of implementations of deep domain adaptation algorithms
MIT License
889 stars 205 forks source link

Question about the model selection #6

Closed taylover-pei closed 5 years ago

taylover-pei commented 5 years ago

Your have done a great job. But I have a question about the training process. In the domain adapation scenario, you tune the parameters of the model according to the validation set of source domain or the target domain? That means, how do you select the best model? According to the evaluation metrics of source domain validation set or the target domain validation set. Thank you.

easezyc commented 5 years ago

Actually, I consider it is a common problem for domain adaptation community. There are only training and test sets for such as Office31, ImageCLEF, OfficeHome datasets. Hence, I use the test set as unlabeled data for aligning distributions and also select the best models according to the performance on the test set.