Skip to content

Latest commit

 

History

History
18 lines (8 loc) · 443 Bytes

File metadata and controls

18 lines (8 loc) · 443 Bytes

CNN Famaliy: implementation based on tensorflow and keras

#CNN-LeNet-5-Based-on-Tensorflow-and-keras

Build CNN leNet-5 net based on Tensorflow and keras FOR BEGINNERS

L2 Regularization, Dropout and expenotial average techs are used for optimise the network.

Based on the MINST data set, the accuracy is 99.4%.

#AlexNet_Tensorflow&Keras

Build AlexNet based on tf and keras. The model can be used as templates for futher application