-
楼主你好:
我这边在本地跑测试1080t batch size设置为4 因为mem不够 size还是456没有减少,这个运行结果可能还有其他影响因素么?
另外这个batch size的设置是最优设置是16么? 理论依据楼主能麻烦贴个链接么?
多谢
![运行结果截图](https://user-images.githubusercontent.com/11207771/70678529-57…
-
Hi and thanks for this awesome repo.
I just checked the original TensorFlow implementation and found a part different from them. In the original implementation. There is a probability of applying …
ghost updated
3 years ago
-
Hello, when I run the program, an error occurred "AttributeError: Can 't pickle local object' SubPolicy. Just set the < locals >. < lambda > '", don't know if you Can help me to solve it? Thank you ve…
-
> The code is tested on Python 2.7 and Tensorflow 1.13.
Why use Python 2 for this and not Python 3? Python 2 is less than six months away from expiration. No one prefers to use Python 2 anymore.
-
Thanks for the great work and code!
I want to reproduce the same augmentation policy(called genotype in your code) with the provided searching code. I folllow the description in ReadME.md and search …
-
## 論文リンク
- [CVF](https://openaccess.thecvf.com/content/CVPR2021/html/Dabouei_SuperMix_Supervising_the_Mixing_Data_Augmentation_CVPR_2021_paper.html)
## 公開日(yyyy/mm/dd)
CVPR 2021
## 概要
## Te…
-
## 一言でいうと
自動で最適なData Augmentationを探索するAutoAugment(#764 )の計算時間を1000倍高速にしたという研究。Data Augmentationの適用もパラメーターの一種と考え、進化戦略(PBA)を用いて良好な結果を出したモデル/Augmentationを残していく形を取っている。モデルのパラメーターが持ち越されるため再計算の必要がない
!…
-
The torchvision transforms now have 2 backends (PIL and Tensor), here are some functional mismatch between them and some may-be-useful features that neither of them support. Details are listed in [tra…
-
Hi, I have a few questions about the training of PyramidNET
1. How long did you take to fully train pyramidNET with horovod (1 node, 4 GPUs?)
2. The version of pytorch and horovod?
3. Any tricks wh…
-
### 論文へのリンク
[[arXiv:2001.04086] GridMask Data Augmentation](https://arxiv.org/abs/2001.04086)
### 著者・所属機関
Pengguang Chen, Shu Liu, Hengshuang Zhao, Jiaya Jia
### 投稿日時(YYYY-MM-DD)
- The …