...
| Issue | PRs | Contributor(s) | Notes |
|---|---|---|---|
| AMP | https://github.com/apache/incubator-mxnet/pull/14173 | @ptrendx | To move foreard: https://github.com/apache/incubator-mxnet/pull/14173#pullrequestreview-235846341 |
| MKLDNN Quantization | https://github.com/apache/incubator-mxnet/pull/14819 | @ZhennanQin | |
| FP32 optimization | #14914, #14818, #14893, #14877, #14783 | @ciyongch, @TaoLv, @juliusshufa, @yinghu5, @TaoLv | |
| bug fix for infer shape partial | #14869 | @royweilai | under review |
| dependency upgrade for mxnet | #14950, #14887, #14588 | @stu1130 | update to latest cuda 10.1 and cudnn 7.5.1 for mxnet and mxnet CI |
| Conversion from FP32 to Mixed Precision Models | https://github.com/apache/incubator-mxnet/issues/14584 | @anirudh2290 | Depend on AMP PR |
MKLDNN RNN Inference Integration(fp32 LSTM and vRNN with tanh and relu) | #14713 | @lihaofd, Tao Lv, @pengzhao-intel | Improves performance of certain operators used in RNN models |
Release Timeline
Following timeline is based on everything goes well.(Added some buffer time)
...