This is an automated email from the ASF dual-hosted git repository.
apeforest pushed a change to branch benchmark
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git.
discard 2f90294 adding API doc for Lamb Phase 1 and 2
discard eadd5de fixing base lamb optimizer
discard 259dd7b initial commit lamb optimizer
add e7e6532 Add fast implementation of LARS (#16122)
add d4bd7cf fix atol for test_preloaded_multi_sgd (#16356)
This update added new revisions after undoing existing revisions.
That is to say, some revisions that were in the old version of the
branch are not in the new version. This situation occurs
when a user --force pushes a change and generates a repository
containing something like this:
* -- * -- B -- O -- O -- O (2f90294)
\
N -- N -- N refs/heads/benchmark (d4bd7cf)
You should already have received notification emails for all of the O
revisions, and so the following emails describe only the N revisions
from the common base, B.
Any revisions marked "omit" are not gone; other references still
refer to them. Any revisions marked "discard" are gone forever.
No new revisions were added by this update.
Summary of changes:
python/mxnet/optimizer/optimizer.py | 317 ++++++++++++++++----
src/operator/contrib/multi_lars-inl.h | 100 +++++++
src/operator/contrib/multi_lars.cc | 56 ++++
.../multi_lars.cu} | 12 +-
src/operator/contrib/multi_sum_sq-inl.h | 100 +++++++
src/operator/contrib/multi_sum_sq.cc | 84 ++++++
src/operator/contrib/multi_sum_sq.cu | 173 +++++++++++
src/operator/contrib/preloaded_multi_sgd-inl.h | 332 +++++++++++++++++++++
src/operator/contrib/preloaded_multi_sgd.cc | 245 +++++++++++++++
.../{amp_graph_pass.cc => preloaded_multi_sgd.cu} | 44 +--
src/operator/optimizer_op-inl.h | 186 ------------
src/operator/optimizer_op.cc | 77 -----
src/operator/optimizer_op.cu | 7 -
.../lenet5_train.py => nightly/test_optimizer.py} | 67 +++--
tests/python/gpu/test_operator_gpu.py | 153 ++++++++++
tests/python/unittest/test_optimizer.py | 73 -----
16 files changed, 1565 insertions(+), 461 deletions(-)
create mode 100644 src/operator/contrib/multi_lars-inl.h
create mode 100644 src/operator/contrib/multi_lars.cc
copy src/operator/{numpy/np_broadcast_reduce_op_index.cu =>
contrib/multi_lars.cu} (77%)
create mode 100644 src/operator/contrib/multi_sum_sq-inl.h
create mode 100644 src/operator/contrib/multi_sum_sq.cc
create mode 100644 src/operator/contrib/multi_sum_sq.cu
create mode 100644 src/operator/contrib/preloaded_multi_sgd-inl.h
create mode 100755 src/operator/contrib/preloaded_multi_sgd.cc
copy src/operator/contrib/{amp_graph_pass.cc => preloaded_multi_sgd.cu} (55%)
copy tests/{python/tensorrt/lenet5_train.py => nightly/test_optimizer.py} (58%)
