Closed aamini closed 7 years ago
Add implementations for two well documented solvers in deep learning: Adagrad and Adadelta. Implementations show increased speedups compared to SGD or even previously implemented adaptive methods in Mocha (Adam).
Thanks a lot!
Add implementations for two well documented solvers in deep learning: Adagrad and Adadelta. Implementations show increased speedups compared to SGD or even previously implemented adaptive methods in Mocha (Adam).