·
2 commits
to main
since this release
First release. Includes:
- fully connected networks up to k layer for multi-class classification and regression
- recurrent neural networks for text generation
- multiple weight initializers such as Xavier and Normal
- weight regularizers such as L2 regularizer
- optimizers such as AdaGrad and SGD
- loss smoother and gradient clippers
- examples in the form of jupyter notebooks on how to build models