Skip to content

[ONNX] PyTorch export to ONNX Opset 7 and 8 - Cont #22421

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Closed
BowenBao wants to merge 17 commits into from
Closed

[ONNX] PyTorch export to ONNX Opset 7 and 8 - Cont #22421

BowenBao wants to merge 17 commits into from

Conversation

@BowenBao
Copy link
Collaborator

@BowenBao BowenBao commented Jul 1, 2019

This is an extension to the original PR #21765

  1. Increase the coverage of different opsets support, comments, and blacklisting.
  2. Adding backend tests for both caffe2 and onnxruntime on opset 7 and opset 8.
  3. Reusing onnx model tests in caffe2 for onnxruntime.

@pytorchbot pytorchbot added oncall: jit Add this issue/PR to JIT oncall triage queue module: build Build system issues module: onnx Related to torch.onnx labels Jul 1, 2019
BowenBao
BowenBao commented Jul 1, 2019
View reviewed changes
@pytorchbot pytorchbot added the module: lint Issues related to our Python/C++ lint rules (run by Travis) label Jul 1, 2019
@BowenBao
Copy link
Collaborator Author

BowenBao commented Jul 2, 2019

@pytorchbot retest this please

@ifedan ifedan added the triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module label Jul 2, 2019
houseroad
houseroad reviewed Jul 11, 2019
View reviewed changes
Copy link
Member

@houseroad houseroad left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Overall direction is good. Left some comments inline.

houseroad
houseroad reviewed Jul 12, 2019
View reviewed changes
Copy link
Member

@houseroad houseroad left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We are closer, see inline comments

tools/run-clang-tidy-in-ci.sh Outdated
-g"-torch/csrc/jit/export.cpp" \
-g"-torch/csrc/jit/import.cpp" \
-g"-torch/csrc/jit/netdef_converter.cpp" \
-g"-torch/csrc/jit/passes/onnx/cast_constant.cpp" \
Copy link
Member

@houseroad houseroad Jul 12, 2019

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I prefer to let clang tidy cover as many files as possible. So let's try to remove the dependence of onnx header files in cast_constant.cpp file. Since there are only two places, hard coding with comments sounds acceptable.

torch/csrc/jit/passes/onnx/cast_constant.cpp Outdated
@@ -0,0 +1,71 @@
#include <torch/csrc/jit/passes/onnx/cast_constant.h>
Copy link
Member

@houseroad houseroad Jul 12, 2019

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

call this pass cast_all_constant_to_floating? cast_constant sounds too general.

facebook-github-bot
facebook-github-bot reviewed Jul 12, 2019
View reviewed changes
Copy link
Contributor

@facebook-github-bot facebook-github-bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@houseroad has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

houseroad
houseroad approved these changes Jul 12, 2019
View reviewed changes
Copy link
Member

@houseroad houseroad left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks good, let's land it.

@facebook-github-bot
Copy link
Contributor

facebook-github-bot commented Jul 12, 2019

@houseroad merged this pull request in b3147bc.

@houseroad houseroad mentioned this pull request Jul 15, 2019
Closed
@drnikolaev
Copy link
Contributor

drnikolaev commented Jul 16, 2019

@lara-hdr @BowenBao @houseroad
Folks, am I doing something wrong?
When I run ./scripts/onnx/test.sh I get:

test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2BackendEmbed_opset9::test_densenet FAILED                                                                                      [  6%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2BackendEmbed_opset9::test_traced_ints FAILED                                                                                   [ 15%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset7::test__dim_arange FAILED                                                                                        [ 16%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset7::test_densenet FAILED                                                                                           [ 17%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset7::test_dynamic_loop FAILED                                                                                       [ 17%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset7::test_loop FAILED                                                                                               [ 24%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset7::test_resnet FAILED                                                                                             [ 26%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset7::test_while_cond FAILED                                                                                         [ 28%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2BackendEmbed_opset7::test_densenet FAILED                                                                                      [ 29%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2BackendEmbed_opset7::test_traced_ints FAILED                                                                                   [ 39%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset8::test__dim_arange FAILED                                                                                        [ 40%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset8::test_alexnet FAILED                                                                                            [ 40%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset8::test_densenet FAILED                                                                                           [ 41%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset8::test_dynamic_loop FAILED                                                                                       [ 41%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset8::test_loop FAILED                                                                                               [ 47%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset8::test_resnet FAILED                                                                                             [ 50%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset8::test_while_cond FAILED                                                                                         [ 51%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2BackendEmbed_opset8::test_densenet FAILED                                                                                      [ 53%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2BackendEmbed_opset8::test_traced_ints FAILED                                                                                   [ 63%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset10::test__dim_arange FAILED                                                                                       [ 63%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset10::test_alexnet FAILED                                                                                           [ 64%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset10::test_densenet FAILED                                                                                          [ 65%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset10::test_dynamic_loop FAILED                                                                                      [ 65%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset10::test_loop FAILED                                                                                              [ 71%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset10::test_nested_loops FAILED                                                                                      [ 73%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset10::test_resnet FAILED                                                                                            [ 74%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset10::test_topk FAILED                                                                                              [ 74%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset10::test_topk_script FAILED                                                                                       [ 74%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset10::test_while FAILED                                                                                             [ 75%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset10::test_while_cond FAILED                                                                                        [ 75%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2BackendEmbed_opset10::test_densenet FAILED                                                                                     [ 76%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2BackendEmbed_opset10::test_traced_ints FAILED                                                                                  [ 86%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset9::test__dim_arange FAILED                                                                                        [ 87%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset9::test_alexnet FAILED                                                                                            [ 87%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset9::test_densenet FAILED                                                                                           [ 88%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset9::test_dynamic_loop FAILED                                                                                       [ 88%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset9::test_loop FAILED                                                                                               [ 94%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset9::test_nested_loops FAILED                                                                                       [ 97%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset9::test_while FAILED                                                                                              [ 98%]
test/onnx/test_pytorch_onnx_caffe2.py::TestCaffe2Backend_opset9::test_while_cond FAILED

@BowenBao
Copy link
Collaborator Author

BowenBao commented Jul 16, 2019

@drnikolaev I'm not able to repro this locally, could you share the error message that you are seeing?
Please also check if your ONNX version is up to date as the one being submoduled in PyTorch. There were required fixes that went in ONNX after 1.5 release.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@facebook-github-bot facebook-github-bot facebook-github-bot left review comments

@houseroad houseroad houseroad approved these changes

Assignees

No one assigned

Labels

Merged module: build Build system issues module: lint Issues related to our Python/C++ lint rules (run by Travis) module: onnx Related to torch.onnx oncall: jit Add this issue/PR to JIT oncall triage queue open source triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

8 participants

@BowenBao @facebook-github-bot @drnikolaev @houseroad @ezyang @ifedan @pytorchbot @mruberry