You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@mxnet.apache.org by pt...@apache.org on 2019/06/28 21:22:31 UTC
[incubator-mxnet] branch master updated (11e6d45 -> ca565a0)
This is an automated email from the ASF dual-hosted git repository.
ptrendx pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
from 11e6d45 [AMP] Move topk from FP16_FP32_FUNCS to FP32_FUNCS (#15342)
add ca565a0 Conversion from FP32 model to Mixed Precision model (#15118)
No new revisions were added by this update.
Summary of changes:
docs/tutorials/amp/amp_tutorial.md | 42 ++
.../automatic-mixed-precision}/README.md | 21 +-
.../amp_model_conversion.py | 119 ++++++
.../common | 0
include/mxnet/c_api.h | 49 +++
python/mxnet/contrib/amp/amp.py | 353 ++++++++++++++++-
python/mxnet/gluon/parameter.py | 44 ++-
python/mxnet/module/executor_group.py | 8 +
python/mxnet/test_utils.py | 83 ++++
src/c_api/c_api_symbolic.cc | 204 ++++++++++
src/nnvm/amp_infer_unknown.cc | 148 +++++++
src/nnvm/low_precision_pass.cc | 261 +++++++++++++
tests/python/gpu/test_contrib_amp.py | 428 +++++++++++++++++++++
tests/python/tensorrt/test_tensorrt_lenet5.py | 2 +-
tests/python/unittest/test_contrib_amp.py | 85 ----
15 files changed, 1739 insertions(+), 108 deletions(-)
copy {python => example/automatic-mixed-precision}/README.md (51%)
create mode 100644 example/automatic-mixed-precision/amp_model_conversion.py
copy example/{quantization => automatic-mixed-precision}/common (100%)
create mode 100644 src/nnvm/amp_infer_unknown.cc
create mode 100644 src/nnvm/low_precision_pass.cc
create mode 100644 tests/python/gpu/test_contrib_amp.py
delete mode 100644 tests/python/unittest/test_contrib_amp.py