SuLi

hello， I have the same problem too, Could you please tell me how to use my own dataset?(obj or images?)

Hello, I have the same problem as yours about count.Do you have find a method for this? And also , I have a problem at complete_scan.py lin325 (ValueError: Can't load...

well,thank you for your reply. And now,when I run this code complete_scan.py I only got this result(*********__0__pred.tfrecord),not a obj.what should I do to produce obj? Best wishes！ Thank you!

Well,I found the cause of the problem. Now,I can get the result of obj my email address is [email protected] I hope I can communicate with you if I can.

Hello    I think you can try to use python2.7 to run this.     best wishes! ------------------ 原始邮件 ------------------ 发件人: "santhu937"

how about to # save_pred_sem? it's too long ago that I can't recall the  operation I used. ------------------ 原始邮件 ------------------ 发件人: "santhu937"

yes  It's right  but if #save_pred_sem maybe you can't get the network model. I just use the test datasets,instead of my datasets.  ------------------ 原始邮件 ------------------ 发件人: "santhu937"

maybe github make that  It isn't I write. ------------------ 原始邮件 ------------------ 发件人: "santhu937"

pred_sem.obj is the output generated by the model target_sem.obj is the target model  ------------------ 原始邮件 ------------------ 发件人: "santhu937"

你好，我是ucas的研究生，很关注您的工作，想了解下您是否愿意提供训练代码？想尝试一些想法。