[GitHub] szha commented on issue #14208: Add support for fast variable-length LSTM

szha commented on issue #14208: Add support for fast variable-length LSTM URL: https://github.com/apache/incubator-mxnet/pull/14208#issuecomment-466183586 @stephenrawls the length array is generally small so I think it's fine to do a copy inside the operator. @reminisce @junrushao1994 are l

[GitHub] szha commented on issue #14208: Add support for fast variable-length LSTM

szha commented on issue #14208: Add support for fast variable-length LSTM URL: https://github.com/apache/incubator-mxnet/pull/14208#issuecomment-465870652 @stephenrawls for now you can do a `CHECK(!use_seq_len)` in CPU kernel. @pengzhao-intel @TaoLv will see how to add it for CPU kernels.

[GitHub] szha commented on issue #14208: Add support for fast variable-length LSTM

szha commented on issue #14208: Add support for fast variable-length LSTM URL: https://github.com/apache/incubator-mxnet/pull/14208#issuecomment-465424858 @stephenrawls swapaxis is doing an actual transpose This is an automate