Nguyen Nguyen Anh

I train all over from scratch ... now. Haiizzz...!

I double the D network... Let's see. It is not going anywhere, darn :( !!!! Steve

> I have never tried training conditional SG2. I think your dataset is too perplexed for such a small (fmaps=0.5) network. I'd try to scale it up to 100-200M parameters...

Wastes of time ! This network only learn portrait-dataset. Weak. Steve

Let's build your own bros.... RetinaNet Mobile/ResNet will do. We need much better accuracy.

StyleGAN current plus ADA whatever can only work with very portrait and focused-simple to detect dataset (faces). Do NOT expect any good results at all with other perplexed dataset. It...

Lưu ý: nên dùng ES 7.3.1, các version cũ crash loạn xạ với Java core dump ! sudo su apt-get update -y apt-get upgrade -y apt-get install build-essential cmake unzip...

Hi, How is it possible that my pre-trained score FiD was 59 in original StyleGanV2-ADA, and in the first 200 kimg with your new codes, it went down to 10...

Here we go at 24,000 kimg (24 million images went through network). - Total images 47,000 (256x256) with 19 conditional classes - 4xV100 batch 64. - RAM/CPU much more than...

> I think the generated images do not look very bad :) So maybe it's just because the dataset is quite challenging, and state-of-the-art GAN models are still limited in...