A Qustion about regular_train_op ----main.py line 66

[DanyaXu]

After reading your article, I was deeply inspired, but there was a problem when reading the code. During training, the optimizer ‘dann_trian_op’ was used, but the optimizer ‘regular_train_op’ was not used. Why?

[daoyuan98]

Hi, the regular_train_op in line66 is just an op for testing when I was writing the code which optimizes only one classification loss and no domain adversarial loss is optimized. While dann_train_op in line67 optimizes the sum of both domain adversarial loss and regular classification loss, which is actually the loss proposed in the paper.

[DanyaXu]

Hi, the regular_train_op in line66 is just an op for testing when I was writing the code which optimizes only one classification loss and no domain adversarial loss is optimized. While dann_train_op in line67 optimizes the sum of both domain adversarial loss and regular classification loss, which is actually the loss proposed in the paper.

Okay, thank you very much for your reply.