Description

Adding tests for operators that can currently be exported/imported from/to ONNX. Correcting the name of some tests. Added ONNX export of LogSoftMax too.

Checklist

Essentials

Please feel free to remove inapplicable items for your PR.

• The PR title starts with [MXNET-$JIRA_ID], where $JIRA_ID refers to the relevant JIRA issue created (except PRs with tiny changes)
• Changes are complete (i.e. I finished coding on this PR)
• All changes have test coverage:
• Unit tests are added for small changes to verify correctness (e.g. adding a new operator)
• Nightly tests are added for complicated/long-running ones (e.g. changing distributed kvstore)
• Build tests will be added for build configuration changes (e.g. adding a new build option with NCCL)
• Code is well-documented:
• For user-facing API changes, API doc string has been updated.
• For new C++ functions in header files, their functionalities and arguments are documented.
• For new examples, README.md is added to explain the what the example does, the source of the dataset, expected performance on test set and reference to the original paper if applicable
• Check the API doc at http://mxnet-ci-doc.s3-accelerate.dualstack.amazonaws.com/PR-$PR_ID/$BUILD_ID/index.html
• To the my best knowledge, examples are either not affected by this change, or have been fixed to be compatible with this change

Changes

ONNX export:

• All matmul tests have been enabled
• test_softplus, test_lrn, test_dropout, test_unsqueeze was missing before - added now
• enabled tests which have the string 'broadcast' in them
• Add default test for flatten
• test_log_ covers onnx operator test test_log*, PyTorch Converted model test test_log_softmax*
• test_elu, test_leakyrelu, test_relu, test_tanh all have onnx operator tests - removing from basic model tests

ONNX import:

• there is no test "test_random" - removed this
• All matmul tests have been enabled
• test_dropout, test_gather, test_log, test_leakyrelu, test_elu, test_max have been enabled
• test_elu, test_leakyrelu, test_relu, test_tanh, test_sigmoid all have onnx operator tests - removing from basic model tests
• Add default test for flatten
• test_log_ covers onnx operator test test_log*, PyTorch Converted model test test_log_softmax*

Comments

• Executed onnx_backend_test, mxnet_backend_test, gluon_backend_test
• ONNX operator tests for test_logsoftmax* is giving a different result compared to MXNet leading to test failure - will verify this manually and enable in a later PR