State-of-the-Art Image Generative Models
I have aggregated some of the SotA image generative models released recently, with short summaries, visualizations and comments. The overall development is summarized, and the future trends are speculated. Many of the statements and the results here are easily applicable to other non-textual modalities, such as audio and video. Summary: The papers we featured in this post belong to either of the f
Large Scale Adversarial Representation Learning
Adversarially trained generative models (GANs) have recently achieved compelling image synthesis results. But despite early successes in using GANs for unsupervised representation learning, they have since been superseded by approaches based on self-supervision. In this work we show that progress in image generation quality translates to substantially improved representation learning performance.
Making Anime Faces With StyleGAN · Gwern.net
A tutorial explaining how to train and generate high-quality anime faces with StyleGAN 1+2 neural networks, and tips/scripts for effective StyleGAN use. Generative neural networks, such as GANs, have struggled for years to generate decent-quality anime faces, despite their great success with photographic imagery such as real human faces. The task has now been effectively solved, for anime faces as
This Waifu Does Not Exist · Gwern.net
I describe how I made the website ThisWaifuDoesNotExist.net (TWDNE) for displaying random anime faces generated by StyleGAN neural networks, and how it went viral. Generating high-quality anime faces has long been a task neural networks struggled with. The invention of StyleGAN in 2018 has effectively solved this task and I have trained a StyleGAN model which can generate high-quality anime faces
Danbooru2021: A Large-Scale Crowdsourced & Tagged Anime Illustration Dataset · Gwern.net
Danbooru2021: A Large-Scale Crowdsourced & Tagged Anime Illustration Dataset Danbooru2021 is a large-scale anime image database with 4.9m+ images annotated with 162m+ tags; it can be useful for machine learning purposes such as image recognition and generation. Deep learning for computer revision relies on large annotated datasets. Classification/categorization has benefited from the creation of I
Large Scale GAN Training for High Fidelity Natural Image Synthesis(BigGAN)::10/6 - 進捗だめです
https://arxiv.org/pdf/1809.11096.pdf -1. AbstractのAbstract 緻密に色々試したら，つぎは計算機で殴ろうな！！ 0. Abstract いわゆる，BigGANという名前で，少し話題になっているGANです．DeepMindのインターンの成果物らしく自己アテンションを組み込んだSelf-Attention GANをベースに様々な最適化手法が取り込まれているようです．Appendix Gにあるように「こうしたら悪い結果が出た」というのを掲載しているので実装等の際には非常に有用となるかと思います．チャンネル数とバッチサイズの大小と精度についても述べられており，双方ともに上げた方が良い結果が望めるとのこと(やはり計算資源のデカさこそ正義らしい)．潜在変数$z$のサンプリング範囲をあらかじめ決めておいて，その範囲外のサンプルを切り捨てる"Trans
IntroVAE: Introspective Variational Autoencoders for Photographic Image Synthesis
We present a novel introspective variational autoencoder (IntroVAE) model for synthesizing high-resolution photographic images. IntroVAE is capable of self-evaluating the quality of its generated samples and improving itself accordingly. Its inference and generator models are jointly trained in an introspective way. On one hand, the generator is required to reconstruct the input images from the no
Inception score
～ただのメモ置き場。なので、内容の正しさについての保証は一切ありませんよ～ copyright (c) 2006-2024 bluewidz, all rights reserved. はじめに Inception score [1]を計算します。このスコアは、GAN (Generative Adversarial Network)が生成した画像の評価値として使われることがあります。 [1]の著者らによるTensorFlow版のコードが[2]にあります。Chainer版が[3]にあります。ここではKerasで試します。 Inception score Inception scoreは、Inceptionモデルで識別しやすい画像であるほど、かつ、識別されるラベルのバリエーションが豊富であるほどスコアが高くなるように設計されたスコアです。 スコアは次のように計算します。\(x_i\)を\(i\
櫟井唯ちゃんの画像を無限に生成する話りぴーと - ｼﾞｮｲｼﾞｮｲｼﾞｮｲ
joisino.hatenablog.com ▲昔の記事 前回の試みから二年以上経ちましたがまだゆゆ式二期は発表されません。(*1) 二期が発表されないためこの記事のタイトルものんのんびよりさんから拝借することになりました。 やはり今話題のディープラーニングでなんとかするしかなさそうです。 三行で説明して Progressive Growing GAN (Tero Karras et. al. 2017) を chainer で実装して櫟井唯さん（ゆゆ式）の画像を生成しました。 こんな感じのができました。（下の方にスクロールするともっとたくさんあります。） github.com レポジトリです。 Progressive Growing GAN とは 浅いネットワークで小さい画像を生成・識別することからはじめ、段階的にネットワークと画像を大きくしていく手法です。太古に流行った AE の事前学
On Convergence and Stability of GANs
We propose studying GAN training dynamics as regret minimization, which is in contrast to the popular view that there is consistent minimization of a divergence between real and generated distributions. We analyze the convergence of GAN training from this new point of view to understand why mode collapse happens. We hypothesize the existence of undesirable local equilibria in this non-convex game
Create Anime Characters with A.I. !
Create Anime Characters With GirlsM anifold Create Anime Characters with A.I. ! Yanghua Jin School of Computer Science Fudan University yanghuajin94@gmail.com Jiakai Zhang School of Computer Science Carnegie Mellon University zhangjk95@gmail.com Minjun Li School of Computer Science Fudan Univerisity li.akizora@gmail.com Yingtao Tian Department of Computer Science Stony Brook University alan.yt.ti
Adversarial Generation of Natural Language
Generative Adversarial Networks (GANs) have gathered a lot of attention from the computer vision community, yielding impressive results for image generation. Advances in the adversarial generation of natural language from noise however are not commensurate with the progress made in generating images, and still lag far behind likelihood based methods. In this paper, we take a step towards generatin
リリース、障害情報などのサービスのお知らせ
最新の人気エントリーの配信
j次のブックマーク
k前のブックマーク
lあとで読む
eコメント一覧を開く
oページを開く
Nvidia says its AI can fix some of the biggest problems in video calls
Top-k Training of GANs: Improving GAN Performance by Throwing Away Bad Samples
lyrn.ai
Danbooru: Anime Image Board
[1810.12686] Evaluating Text GANs as Language Models
[1810.06758] Discriminator Rejection Sampling
https://research.nvidia.com/sites/default/files/pubs/2017-10_Progressive-Growing-of/karras2018iclr-paper.pdf
GitHub - hwalsuklee/tensorflow-generative-model-collections: Collection of generative models in Tensorflow
