|
Post by hagiasora on May 2, 2018 7:17:38 GMT
i thought that ADAM optimizer is not always the best optimizer for every data in deep learning, but most people use ADAM in their works, is anybody have other opinion?
|
|
|
Post by ucalyptus on May 2, 2018 18:24:35 GMT
I largely depends on your network architecture.Your question seems too much general. Adam is good choice (not best) for a CNN but it all matters with what kind of data you are working on.
|
|
|
Post by deadprez on May 5, 2018 13:37:28 GMT
Depends a lot on your data. For sparse input data, I feel that most of the adaptive learning rate optimizers are going to have similar performance.
|
|