[GitHub] sxjscience commented on issue #9363: Incorrect weight_decay implementation in AdaGrad

sxjscience commented on issue #9363: Incorrect weight_decay implementation in AdaGrad URL: https://github.com/apache/incubator-mxnet/issues/9363#issuecomment-373255756 OK, we need to reopen it then. Get Outlook for iOS Fr

[GitHub] sxjscience commented on issue #9363: Incorrect weight_decay implementation in AdaGrad

sxjscience commented on issue #9363: Incorrect weight_decay implementation in AdaGrad URL: https://github.com/apache/incubator-mxnet/issues/9363#issuecomment-372435894 Closed by the PR. Is the AdaDelta solved? This is an aut

[GitHub] sxjscience commented on issue #9363: Incorrect weight_decay implementation in AdaGrad

sxjscience commented on issue #9363: Incorrect weight_decay implementation in AdaGrad URL: https://github.com/apache/incubator-mxnet/issues/9363#issuecomment-358345278 For SGD, not merging in the wd term will linearly scale the ?real? wd and the impact is smaller compared to Adagrad. Howe