# mlart stylegan2ada machinelearning generativeart raredigitalart digitalpainting cryptoart eth watercolour ai artificialintelligence

AIによる自動生成（StyleGAN2を活用したとのこと）

顔だけじゃなくて、体も生成されてて凄い。なかには漫画っぽいのも生成されてて、言葉らしきものまで生成されているが、言語は解読不能
https://t.co/sTU6rSDVvv

2223 4093
2021-01-20

すべて生成されたイラスト．手や体を切り抜いた画像のデータセットでデータ拡張してStyleGAN2を訓練したみたい

19 62
2021-01-19

-- this anime isn't real --

(teaser for the stylegan2 model that is to be released soon™ ... keep an eye on and )

78 302
2021-01-18

StyleGAN2-generated anime-style images (thread)

31 130
2020-12-06

機械学習させたFurryの画像のlatent space walkです。ぶっちゃけ怖い。そんで750%のUpresしてます（もとは256x256px --> 1536x1536px）So I returned to my AI furry StyleGan2. It's friggin scary. I then used AI to upres the whole thing from 256px to 1536px. Not bad!

1 6
2020-11-27

stylegan2-ada anime n=1k finetuning

2 24
2020-10-10

stylegan2 cartoonset is now on runwayml
link: https://t.co/z12PR034qy

7 54
2020-09-25

Following up, another interpolation video of the same model much later in training. It really seems that, despite this one being trained on random cropping, StyleGAN2 prefers generating centrally-placed objects (especially evident in comparison with BigGAN...) https://t.co/XQwTEb7NbY

4 25
2020-05-13

AI generated (StyleGAN2 still) + Waifu2x denoising and upscaling. OwO

3 43
2020-05-04

Hitting the limits of StyleGAN2's ability to represent complex structures (trained on Danbooru2019)

9 56
2020-04-13

Pokemon now at 540 kimg (~12.5 hours) of training

8 43
2020-04-13

Seven hours (300 kimg) into training a pokemon model from scratch

16 118
2020-04-12

A gorgeous bunch of watercolours generated using Stylegan2. Trying to pick the best from these to create the videos for the "Ode" series: https://t.co/e0TIU2BnSS


0 4
2020-04-04

A journey through conditional StyleGAN2 label space.
The z vector is fixed, and only the labels are changing, based on weighted averages of the w vectors produced after mapping the doc2vec encodings of the text at the bottom of the video.

4 28
2020-03-27

stylegan2 + ConSinGAN + CAIN

0 29
2020-03-26

StyleGAN2を使った自動生成の続きで、今回はDanbooru2018データセットを試してみました。以下から顔部分の切り抜き済みデータをお借りしています。
https://t.co/VmfudvpqGK
学習時間は前回と同じくローカルGPUで20時間程度。溶けている部分や過学習っぽいキャラもいるのでまだ調整は必要そうですね。

33 88
2020-03-11

More stochastic weight average examples. The middle image shows the average of the left and right. Script (swa .py) is here in Peter's StyleGAN2 fork: https://t.co/4tMan2xx4M

credits and - Thanks for sharing!

2 16
2020-01-30

Cursed StyleGAN2 people after a deep style transfer, makes some interesting things

26 136
2020-01-22