question about result

[yandou904]

Hello!
I'm sorry to bother you.I have run the code and tried to adjust the parameters base_lr,but I found that the experimental results were quite different from the paper.The picture below is the best result after I adjusted the parameters.Can you tell me about other ways to improve performance?Thanks!I am looking forward to your reply.

[chaoshangcs]

Hello! I'm sorry to bother you.I have run the code and tried to adjust the parameters base_lr,but I found that the experimental results were quite different from the paper.The picture below is the best result after I adjusted the parameters.Can you tell me about other ways to improve performance?Thanks!I am looking forward to your reply.

[chaoshangcs]

Hi, thanks for your message. First, please note that we have updated the results in table 4, 5 and 6 with corresponding explanation in the appendix. Your results are very similar the new tables. Second, we upgraded our implementation after receiving many valuable feedbacks. The hyperparameters might be different. Tuning the lr and lr steps will be really helpful. If you have any question, please let me know. Thanks.

[zwzhang123]