Train with 1000

Ranking

Accuracy ▼	Cross entropy ▼	Name ▼	Code ▼	Date ▼	Paper ▼
0.9786	0.1396	simple cnn minst	https://github.com/k...	2019/02/27
0.9458	0.2897	Train1000 wig ensemble	https://github.com/m...	2018/11/11	[01]
0.9364	0.4976	Train1000 sample	https://github.com/m...	2018/11/09

[01]	M. Tanaka, Weighted Sigmoid Gate Unit for an Activation Function of Deep Neural Network, arXiv preprint arXiv:1810.01829, 2018.

The reference code is here, which includes obtaining 1,000 samples and sample network training for each dataset.

Please submit your results of train with 1000!