You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@mxnet.apache.org by bg...@apache.org on 2022/02/16 07:34:47 UTC
[incubator-mxnet] branch master updated (4d915f7 -> f35a4d7)
This is an automated email from the ASF dual-hosted git repository.
bgawrych pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
from 4d915f7 Improve MaskedSoftmax by oneDNN (#20853)
add f35a4d7 [FEATURE] Add quantized version of reshape with DNNL reorder primitive. (#20835)
No new revisions were added by this update.
Summary of changes:
python/mxnet/amp/lists/symbol_fp16.py | 2 +
src/operator/nn/dnnl/dnnl_reshape.cc | 1 +
src/operator/numpy/np_matrix_op-inl.h | 4 +
...ntized_flatten.cc => dnnl_quantized_reshape.cc} | 58 +++++----
src/operator/quantization/quantized_reshape-inl.h | 73 ++++++++++++
src/operator/quantization/quantized_reshape.cc | 132 +++++++++++++++++++++
tests/python/dnnl/subgraphs/test_conv_subgraph.py | 28 +++++
tests/python/dnnl/subgraphs/test_fc_subgraph.py | 29 ++++-
tests/python/quantization/test_quantization.py | 40 +++++++
9 files changed, 341 insertions(+), 26 deletions(-)
copy src/operator/quantization/dnnl/{dnnl_quantized_flatten.cc => dnnl_quantized_reshape.cc} (51%)
create mode 100644 src/operator/quantization/quantized_reshape-inl.h
create mode 100644 src/operator/quantization/quantized_reshape.cc