Osman Tursun

@Turlan my curve is similar to you. I think it is because discriminator is learning in the beginning.

Hi, I don't know which one is better, but edge-connect network design should be efficient than deepfill v2, and I think using gated convolution is not so critical since dilated...

@jspitz is there any tutorial for `script` option?

@yannis1962 Thanks for your help. Overall, it is working. However, there is a bug related to " ' " (hamza). Is it because it is a one-to-one mapping ? For...

I also have the same problem. Deep all performance reaches 80%.

@silvia1993 I think I made a mistake. I was looking at the results in the terminal. Should I check the results in the tflog? did you report val_test result or...

deep all with simple data augmentation, smaller batch size with more epochs will beat the JigenDG by 2 or 3 percent. On Wed, Jan 6, 2021 at 12:57 PM Rui...

Hi, there is no regularised mapping table for the pinyin-based Latin alphabet. However, if you are interested you can read this paper: [Noisy_Uyghur_Text_Normalization](https://www.researchgate.net/publication/320279773_Noisy_Uyghur_Text_Normalization)

@umdreamer The calculation is correct, but you can check this one as well: https://github.com/qliu24/SAKE

The following setup works for me. You can read more [here](https://bugs.launchpad.net/ubuntu/+source/evolution/+bug/1866974) ```sudo mkdir /etc/gnutls cat