-
## 一言でいうと
テキスト生成において品質と多様性という2面で評価を行う方法と、その評価において基本的な最尤推定のモデル(MLE)がGANベースよりも優位であることを示した研究。softmaxのtemperatureは高いと確率が等しい=多様性に寄与、低いと決定的=品質に寄与するので、これを変化させた時のカーブで比較する。
![image](https://user-images.g…
-
## 一言でいうと
GANの学習は、GとDの間にナッシュ均衡を見つける問題と見立てられる。これを見つけるにはsimultaneousなGDが用いられるが、両者の勾配ベクトルが特定の固有値を持つ場合収束しない。そこで正則化項を導入しこれを防いでいる。
### 論文リンク
https://arxiv.org/abs/1705.10461
### 著者/所属機関
Lars …
-
### 論文へのリンク
[[arXiv:2006.14567] Taming GANs with Lookahead](https://arxiv.org/abs/2006.14567)
### 著者・所属機関
Tatjana Chavdarova, Matteo Pagliardini, Martin Jaggi, Francois Fleuret
- Idiap, EP…
-
The progressive growing GAN paper from NVIDIA is probably the breakout GAN paper of the year. It would be cool to have a DeepChem implementation.
http://research.nvidia.com/publication/2017-10_Prog…
-
Hi Pushkar,
It 's the same guy who contacted you on linkedin. All over the internet we have seen PATE-GANs being used for tabular data synthetic generation. We wanted to know it's feasibility for i…
-
I've been using TorchGAN to train the global generator/discriminator of pix2pixHD and got great results. I wonder if it's supported to load the pre-trained global models, add local enhancer layers on …
-
I'm a bit confused, for the "previous" repository progressivegan3d there was a clear way to train a GAN using a single python call. Now I get the impression that I need to copy a rather large chunk of…
-
-
Hi!
I am implementing GANs inversion using repo [encoder4editing](https://github.com/omertov/encoder4editing).
But I saw that they trained the model with stylegan2 generator to be fixed. In your n…
-
Generative adversarial networks description
mhrmm updated
4 years ago