zixuanweeei commented on issue #15497: [MKLDNN] Independent gradients requests
check with respect to weights and bias of convolution
URL: https://github.com/apache/incubator-mxnet/pull/15497#issuecomment-511638619
@ciyongch The equality assertion error happened with the outputs of
convolut
zixuanweeei commented on issue #15497: [MKLDNN] Independent gradients requests
check with respect to weights and bias of convolution
URL: https://github.com/apache/incubator-mxnet/pull/15497#issuecomment-511187153
I will take some tests to see whether a low bar works.
-
zixuanweeei commented on issue #15497: [MKLDNN] Independent gradients requests
check with respect to weights and bias of convolution
URL: https://github.com/apache/incubator-mxnet/pull/15497#issuecomment-511183395
@pengzhao-intel @TaoLv Please re-review on this PR. It should be noted that
zixuanweeei commented on issue #15497: [MKLDNN] Independent gradients requests
check with respect to weights and bias of convolution
URL: https://github.com/apache/incubator-mxnet/pull/15497#issuecomment-510794204
@matteosal Thanks. That's great. I am working on a unit test for this
featur
zixuanweeei commented on issue #15497: [MKLDNN] Independent gradients requests
check with respect to weights and bias of convolution
URL: https://github.com/apache/incubator-mxnet/pull/15497#issuecomment-510705077
@matteosal I tested your example with commit 9ca042, and it ran successfully