Skip to content

Commit 23b1ea0

Browse files
authored
Create README.md
1 parent bad8a0b commit 23b1ea0

File tree

1 file changed

+30
-0
lines changed

1 file changed

+30
-0
lines changed

README.md

+30
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,30 @@
1+
# Shake Drop regularization
2+
PyTorch implementation of shake-drop regularization.
3+
Author Torch implementations is [here](https://github.com/imenurok/ShakeDrop).
4+
5+
## Dependencies
6+
- python 3.5+
7+
- PyTorch 1.0.0
8+
9+
## Accuracy
10+
11+
### CIFAR-100
12+
|Model|Method|Level|Alpha|Beta|This implementaion |Paper|
13+
|:---:|:---:|:---:|:---:|:---:|:---:|:---:|
14+
|PyramidNet|ShakeDrop|Batch|[-1, 1]|[0, 1]|83.90|83.78|
15+
16+
![CIFAR-100](cifar100.png)
17+
18+
## Train PyramidNet(depth=110, alpha=270) with shake-drop for CIFAR-100
19+
```bash
20+
python train.py --epochs 300 --batch_size 128 --label 100 --lr 0.5 --depth 110 --alpha 270 --snapshot_interval 10
21+
```
22+
23+
## References
24+
Yoshihiro Yamada, Masakazu Iwamura, Koichi Kise. "ShakeDrop regularization"
25+
[ICLR2017 OpenReview](https://openreview.net/forum?id=S1NHaMW0b)
26+
27+
Yoshihiro Yamada, Masakazu Iwamura, Takuya Akiba, Koichi Kise. "
28+
ShakeDrop Regularization for Deep Residual Learning"
29+
[arXiv:1802.02375v2](https://arxiv.org/abs/1802.02375)
30+

0 commit comments

Comments
 (0)