A Style-Based Generator Architecture for Generative Adversarial Networks
推荐理由
The latest generator in this list has very powerful latent spaces, including approximately accurate 3D rotations.
说明
这是2019年由Tero Karras FI发布的一部演示,主题归在LLMs & generative AI。它讲的是generative adversarial networks这一支里的具体问题,而不是把整个领域扫一遍。适合已经能看懂基本术语、想继续往下走的人。内容上可以概括成:A short demo video for NVIDIA's StyleGAN paper, which proposes a style-based generator for GANs. It shows face synthesis with style mixing and scale-specific control over attributes like pose and freckles.这些读者给出的理由,核心是:Cited as an example of how realistic GAN-generated faces had become, and of what a well-structured latent space can do.它贴着当时的技术,看的时候要记得那是那个年份的判断。从2018年到2019年,有0位不同的读者在0个不同月份把它推荐出来,其中明确叫别人去看的有0次。适合先看结果再决定要不要深挖的人。先看它做成了什么。如果已经会了基本定义,就把注意力放在它怎么把generative adversarial networks变成能跟着复述的例子。
视频信息
- 频道
- Tero Karras FI
- 时长
- 6 分 18 秒
- 发布年份
- 2019