-
논문 제목 : TryOnDiffusion: A Tale of Two UNets
분야 : cv
논문 링크 : https://arxiv.org/pdf/2306.08276.pdf
발표 자료 : https://www.notion.so/hama-lab/TryOnDiffusion-A-Tale-of-Two-UNets-5c36f56b58084fdabc310078bd…
-
## Why
Machine Learning 輪講は最新の技術や論文を追うことで、エンジニアが「技術で解決できること」のレベルをあげていくことを目的にした会です。
prev. #9
## What
話したいことがある人はここにコメントしましょう！
面白いものを見つけた時点でとりあえず話すという宣言だけでもしましょう！
-
The new paper ["Image Augmentations for GAN Training"](https://arxiv.org/abs/2006.02595#google), Zhao et al 2020b, reports:
> Data augmentations have been widely studied to improve the accuracy and…
gwern updated
3 years ago
-
**Is your feature request related to a problem? Please describe.**
This project implements a Generative Adversarial Networks framework (GAN) to generate 3D voxel data. The goal is to train a GAN to p…
-
## Project Request
"SentinelAI: Harnessing Satellite Vision for Rapid Disaster Response and Life-saving Insights"
---
| Field | Computer Vision , DeepLearning |
| ------ | ---------------…
-
## Keyword: out of distribution detection
There is no result
## Keyword: out-of-distribution detection
There is no result
## Keyword: expected calibration error
### Soft Augmentation for Image Class…
-
-
## 内容
用mixup为对比学习做数据增强，可以在SimCLR，MoCo，BYOL上用，在图像、语音和表格数据（为啥这类文章都不做文本？）上都有提升。也验证了方法的正则化效果，在（1）数据不够（2）数据增强的领域知识不够，两种情况下明显改善了对比学习。
具体就是为batch中每个instance分配virtual label（其实就是它的序号对应的one-hot），然后再mixup生成新…
-
We need to understand the backend architecture so that we can be confident about it.
-Model1: GAN
We have generator & discriminator.
Based on the user's choice it learns & improves upon.
It…
-
## What's the problem?
Many Wordplay examples need text content or data to be interesting -- a hollow algorithm shell without interesting content tends not to inspire. The internet can be a source …