Reduce binary size of libraries consuming ONNX (part 1/2) by pranavsharma · Pull Request #2643 · onnx/onnx

pranavsharma · 2020-03-05T09:40:03Z

Several parts of the op sec like the main op description, attributes, input and output descriptions become part of the binary that consumes ONNX e.g. onnxruntime causing an increase in its size due to strings that take no part in the execution of the model or its verification.

Setting __ONNX_NO_DOC_STRINGS doesn't really help here since (1) it's not used in the SetDoc(string) overload (see https://github.com/onnx/onnx/blob/master/onnx/defs/schema.cc#L444) and (2) it's not used at all for attributes, inputs and outputs. We should do something similar to https://github.com/onnx/onnx/blob/master/onnx/defs/schema.h#L322 to fix these things.

This PR takes care of the SetDoc calls and inputs/outputs. I'll send a separate PR for the attributes since they're a bit more involved and this PR is already big. These and the attribute changes together reduce the binary size of onnxruntime by at least 95k (tested using vs2017) with LTO enabled.

Reference: #2628

…on in the spec.

CLAassistant · 2020-03-05T09:40:09Z

All committers have signed the CLA.

pranavsharma · 2020-03-05T09:43:44Z

    11,
    OpSchema()
-        .SetDoc(std::string(BitShift_ver11_doc) + GenerateBroadcastingDocMul())
+        .SetDoc(GetBitShiftDoc())


The addition of the output from GenerateBroadcastingDocMul() inhibits the linker from throwing away BitShift_ver11_doc. Hence the introduction of the function GetBitShiftDoc().

pranavsharma · 2020-03-05T09:46:04Z

 std::function<void(OpSchema&)> BinaryLogicDocGenerator_opset1(
    const char* name) {
  return [=](OpSchema& schema) {
+#ifndef __ONNX_NO_DOC_STRINGS


This kind of setup of the doc doesn't help in removing these strings from the binary. Hence the check for __ONNX_NO_DOC_STRINGS here.

pranavsharma · 2020-03-05T09:47:18Z

  }

-  OpSchema& SetDoc(std::string doc);
+  OpSchema& SetDoc(const std::string& doc) {


Changed to const-ref. No need to pay for the copy of doc/description if we're not going to copy it. Ditto for Input() and Output() functions.

pranavsharma · 2020-03-05T23:05:01Z

cc @linkerzhang

pranavsharma · 2020-03-05T23:05:16Z

Can this make it to 1.7 release?

pranavsharma · 2020-03-06T07:28:31Z

Is there any existing issue going on with CircleCI? I see that ci/circleci: py3.6-clang7-ubuntu16.04 is failing for other PRs too that were a day or so old.

linkerzhang · 2020-03-09T01:56:48Z

circleci is not a required ci for onnx, which may be ignored. @spandantiwari for the awareness of the failure, btw.

…enience

pranavsharma · 2020-03-18T02:37:43Z

Any more comments that need to be addressed here?

gramalingam · 2020-03-26T03:55:41Z

Unfortunately, there appears to be some conflicts to be resolved.

pranavsharma · 2020-03-28T05:02:52Z

Can this be merged once CI is complete?

…o issue_2628

* Fix Greater/LessOrEqual function definition (#2645) * Fix Greater/LessOrEqual function definition * Update test data Co-authored-by: Ke Zhang <kezhan@microsoft.com> * Suppress a warning in unsqueeze (#2637) I keep getting this warning when building PyTorch: ``` In file included from /home/hong/wsrc/pytorch/third_party/onnx/onnx/defs/tensor/utils.h:6, from /home/hong/wsrc/pytorch/third_party/onnx/onnx/defs/tensor/defs.cc:4: /home/hong/wsrc/pytorch/third_party/onnx/onnx/defs/tensor/defs.cc: In lambda function: /home/hong/wsrc/pytorch/third_party/onnx/onnx/defs/tensor/defs.cc:1414:22: warning: unnecessary parentheses in declaration of â��iâ�� [-Wparentheses] for (size_t(i) = 0; i < axes.size(); ++i) { ^ /home/hong/wsrc/pytorch/third_party/onnx/onnx/defs/schema.h:959:12: note: in definition of macro â��ONNX_OPERATOR_SET_SCHEMA_EXâ�� return impl.SetName(#name) \ ^~~~ /home/hong/wsrc/pytorch/third_party/onnx/onnx/defs/tensor/defs.cc:1369:1: note: in expansion of macro â��ONNX_OPERATOR_SET_SCHEMAâ�� ONNX_OPERATOR_SET_SCHEMA( ``` This commit should fix it and modernize the code a bit. Co-authored-by: Ke Zhang <kezhan@microsoft.com> * [Training] Add Adagrad optimizer operator (#1955) * Adagrad draft * MIMO * Support multiple tensors to be optimized * Address comments * Move optimizers to a new place Remove copied Add momentum Save Remove momentum Fix Move constants to attributes * Fix build * Add shape test Add two node tests Update test coverage * Fix shape inf * Fix shape inf * fix shape inf * Format * Add function type * Merge lines * Format * Fix version number * Update op version in model files * Fix a test function and update related test files * Update onnx/backend/test/case/node/adagrad.py * Remove unused file * sync docs * Fix shape test * sync doc * sync with master * Update onnx/defs/training/defs.cc Co-Authored-By: Michał Karzyński <postrational@users.noreply.github.com> * sync doc * address comments * address a minor comment * Polish one line Co-authored-by: Michał Karzyński <postrational@users.noreply.github.com> * [Training] SG with Momentum Optimizer (#1959) * SG with Momentum * Registrate Op Fix Update other docs * Add shape inference code and polish definition * Update docs * Add test cases and fix several bugs * Remove accidently added copy * Alpha -> alpha & Beta -> beta * Clarify an attribute * Fix an attribute * Fix bug * Fix missing attributes * sync doc * Remove unused domain * sync with master Co-authored-by: Chin Huang <chhuang@us.ibm.com> * Change type of label tensor to int32/int64 in SoftmaxCrossEntropyLoss spec. (#2667) * Update Pow input types in Opset 12 (#2666) * Update Pow input types in Opset 12 * gen doc and tests * remove uints and 8 bit ints * add tests * remove uint int x tets * Adding CI for ONNX Debug mode (Linux, OSX) (#2651) * adding an osx build, linux build, with and without onnx_ml for debug mode * test debug mode with ONNX_ML=1 * Rename OPTIONAL to OPTIONAL_VALUE (#2682) Co-authored-by: G. Ramalingam <grama@microsoft.com> * Update Batchnorm test (#2674) * Update Batchnorm test * relax shape inference on scalar * Remove unnecessary copies and std::move (#2684) * Update sequence test case so input is not scalar and splits are specified (#2675) * Update sequence test case to input is not scalar and splits are specified * Add spaces to make the checker happy * Use cmake GNUInstallDirs (#2661) https://cmake.org/cmake/help/latest/module/GNUInstallDirs.html this make allow install the libraries (and headers) in different location than `lib` (Gentoo uses lib64 for 64-bits libs) also change the .cmake files for avoid conclicts if build both 32-bis and 64-bits (avoids conflict/overwrite files) Co-authored-by: Ke Zhang <kezhan@microsoft.com> * Add 'ignore_index' input in the spec for SoftmaxCrossEntropyLoss and NLLLoss. (#2680) * Add 'ignore_index' input in the spec for SoftmaxCrossEntropyLoss and NLLLoss. * Add tests. * build break. * build break. * clean up. * build break. * Change ignore_index to attribute. * Change ignore_index to attribute. * PR feedback. * PR feedback. * Make ignore_index optional in NLLLoss. * Build break. * remove trailing spaces to fix build break. * Build break. * Update spec doc. * Fix NLLLoss function definition to fix test: test_negative_log_likelihood_loss_input_shape_is_NCd1d2_with_weight_reduction_sum_ignore_index_expanded * PR feedback. * Fix test for softmax cross entropy loss to exclude ignored_index'ed weights from the sum of weights. * Build break. * Reduce binary size of libraries consuming ONNX (part 1/2) (#2643) * Change the return type for the zipmap operator to match the description in the spec. * Reduce binary size of libraries consuming ONNX (part 1/2) * Fix build error * Replace separate Get*Doc() functions with easy macro for greater convenience * Add one more macro for complicated operator doc documentation. Co-authored-by: Ke Zhang <kezhan@microsoft.com> * Update pybind (#2340) (#2688) * Change version number for release verification Change version number for release verification Co-authored-by: Takeshi Watanabe <take-cheeze@users.noreply.github.com> Co-authored-by: Ke Zhang <kezhan@microsoft.com> Co-authored-by: Hong Xu <hong@topbug.net> Co-authored-by: Wei-Sheng Chin <wschin@outlook.com> Co-authored-by: Michał Karzyński <postrational@users.noreply.github.com> Co-authored-by: M. Zeeshan Siddiqui <mzs@microsoft.com> Co-authored-by: Lara Haidar <haidar.lara@gmail.com> Co-authored-by: Vinitra Swamy <vinitras@gmail.com> Co-authored-by: Changming Sun <chasun@microsoft.com> Co-authored-by: G. Ramalingam <grama@microsoft.com> Co-authored-by: Changming Sun <me@sunchangming.com> Co-authored-by: Scott McKay <skottmckay@gmail.com> Co-authored-by: Gustavo Alvarez <462213+sl1pkn07@users.noreply.github.com> Co-authored-by: Pranav Sharma <prs@microsoft.com>

snnn · 2020-04-21T04:52:55Z

This change broke raspberrypi build.

snnn · 2020-05-28T17:15:45Z

This change also break the build on ubuntu 14.04.

See: microsoft/onnxruntime#4048

* Change the return type for the zipmap operator to match the description in the spec. * Reduce binary size of libraries consuming ONNX (part 1/2) * Fix build error * Replace separate Get*Doc() functions with easy macro for greater convenience * Add one more macro for complicated operator doc documentation. Co-authored-by: Ke Zhang <kezhan@microsoft.com>

Pranav Sharma and others added 13 commits April 24, 2018 17:33

Change the return type for the zipmap operator to match the descripti…

cd0d2a9

…on in the spec.

Merge branch 'master' into master

4041130

Merge branch 'master' into master

008d39e

Merge remote-tracking branch 'upstream/master'

da32d89

Merge branch 'master' of https://github.com/pranavsharma/onnx

1d2c937

Merge remote-tracking branch 'upstream/master'

d6e025c

Merge remote-tracking branch 'upstream/master'

7ac2b93

Merge remote-tracking branch 'upstream/master'

95555b6

Merge remote-tracking branch 'upstream/master'

7607ed0

Merge remote-tracking branch 'upstream/master'

ab29702

Merge remote-tracking branch 'upstream/master'

04b7df0

Merge branch 'master' of https://github.com/onnx/onnx

db8aadf

Reduce binary size of libraries consuming ONNX (part 1/2)

edbca5e

pranavsharma requested a review from a team as a code owner March 5, 2020 09:40

pranavsharma commented Mar 5, 2020

View reviewed changes

pranavsharma changed the title ~~Reduce binary size of libraries consuming ONNX (part 1/2)~~ WIP: Reduce binary size of libraries consuming ONNX (part 1/2) Mar 5, 2020

Fix build error

4351c90

pranavsharma changed the title ~~WIP: Reduce binary size of libraries consuming ONNX (part 1/2)~~ Reduce binary size of libraries consuming ONNX (part 1/2) Mar 5, 2020

linkerzhang reviewed Mar 9, 2020

View reviewed changes

Comment thread onnx/defs/math/defs.cc Outdated

Replace separate Get*Doc() functions with easy macro for greater conv…

d53d9b6

…enience

linkerzhang reviewed Mar 9, 2020

View reviewed changes

Comment thread onnx/defs/logical/defs.cc

Pranav Sharma added 2 commits March 11, 2020 23:46

Add one more macro for complicated operator doc documentation.

34bfe71

Merge remote-tracking branch 'upstream/master' into issue_2628

9e44148

Pranav Sharma added 2 commits March 18, 2020 18:44

Merge remote-tracking branch 'upstream/master' into issue_2628

59f4ed3

Merge remote-tracking branch 'upstream/master' into issue_2628

4d6ae6a

Merge remote-tracking branch 'upstream/master' into issue_2628

b0efd60

gramalingam approved these changes Mar 26, 2020

View reviewed changes

Merge remote-tracking branch 'upstream/master' into issue_2628

e8dccc8

linkerzhang approved these changes Mar 30, 2020

View reviewed changes

linkerzhang and others added 3 commits March 30, 2020 14:34

Merge branch 'master' into issue_2628

94ade2b

Merge remote-tracking branch 'upstream/master' into issue_2628

da399d0

Merge branch 'issue_2628' of https://github.com/pranavsharma/onnx int…

ac1f6a3

…o issue_2628

gramalingam merged commit 674438c into onnx:master Mar 30, 2020

chinhuang007 added this to the 1.7 milestone Mar 31, 2020

Conversation

pranavsharma commented Mar 5, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

CLAassistant commented Mar 5, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

pranavsharma Mar 5, 2020

Choose a reason for hiding this comment

Uh oh!

pranavsharma Mar 5, 2020

Choose a reason for hiding this comment

Uh oh!

pranavsharma Mar 5, 2020

Choose a reason for hiding this comment

Uh oh!

pranavsharma commented Mar 5, 2020

Uh oh!

pranavsharma commented Mar 5, 2020

Uh oh!

pranavsharma commented Mar 6, 2020

Uh oh!

linkerzhang commented Mar 9, 2020

Uh oh!

Uh oh!

Uh oh!

pranavsharma commented Mar 18, 2020

Uh oh!

gramalingam commented Mar 26, 2020

Uh oh!

pranavsharma commented Mar 28, 2020

Uh oh!

snnn commented Apr 21, 2020

Uh oh!

snnn commented May 28, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

pranavsharma commented Mar 5, 2020 •

edited

Loading

CLAassistant commented Mar 5, 2020 •

edited

Loading

snnn commented May 28, 2020 •

edited

Loading