@DickJC123 @ptrendx I think the new features are very useful, since we are working on large scale NLP models, where most of the models are based on RNN cells. For example, we need cell state clipping and projection clipping to make the models overcome the possible overfitting. Similarly, it would be very helpful if we can have projection in the RNN cell, in that way, we can improve the memory size to significantly improve the model capacity. So, looking forward to the new features in cuDNN v7.2 to be in MXNet.
@pengzhao-intel Also looking forward to CPU support! Thanks. [ Full content available at: https://github.com/apache/incubator-mxnet/issues/9543 ] This message was relayed via gitbox.apache.org for [email protected]
