We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
Merge pull request #1011 from chinhuang007/clarify-tf2onnx-readme Add project clarification to README
Merge pull request #1008 from chinhuang007/support-tf-2.8 Update Tensorflow (2.8.0) and ONNX (1.10.2) versions
support batch normalization opset 15 (#1001) * support batch normalization opset 15 Signed-off-by: degaochu <[email protected]> * skip node test for previous version Signed-off-by: degaochu <[email protected]> * fix np.testing.assert_almost_equal issue Signed-off-by: degaochu <[email protected]>
update Optional operator to return right type value for empty value (#996) Signed-off-by: degaochu <[email protected]>
Merge pull request #993 from chinhuang007/update-tf-version Update Tensorflow Version for Modelzoo Test
Support optional operators (#980) * add optional data type support Signed-off-by: degaochu <[email protected]> * fix seq test model issue Signed-off-by: degaochu <[email protected]> * update support status Signed-off-by: degaochu <[email protected]> * support identity opset 16 Signed-off-by: degaochu <[email protected]> * fix issue Signed-off-by: degaochu <[email protected]> * fix issues for onnx 1.9.0 Signed-off-by: degaochu <[email protected]> * fix attr proto cross version issue Signed-off-by: degaochu <[email protected]> * update opset list and status doc Signed-off-by: degaochu <[email protected]>
Merge pull request #977 from chinhuang007/support-shape-15 Add support for Shape opset 15
Add support for Pow opset 15 (#976) Add support for Pow opset 15, where bfloat16 data type is added Also add a filter in test_onnx_backend.py to exclude tests with optional inputs.
support use user tensorflow graph when training mode is enabled (#968) Signed-off-by: degaochu <[email protected]>
Release 1.9.0 (#964) * release 1.9.0 Signed-off-by: degaochu <[email protected]> * release 1.9.0 Signed-off-by: degaochu <[email protected]> * release 1.9.0 Signed-off-by: degaochu <[email protected]> * release 1.9.0 Signed-off-by: degaochu <[email protected]> * release 1.9.0 Signed-off-by: degaochu <[email protected]> * trigger CLA check Signed-off-by: degaochu <[email protected]>
Merge pull request #961 from chinhuang007/support-gru-14 Add batchwise support for gru, lstm, rnn
Merge pull request #958 from chinhuang007/batchnorm-14 Partially support BatchNormalization opset 14
Add suppor for ONNX model with external data (#957) Add suppor for ONNX model with external data with a new optional command line argument to specify the external data directory. Signed-off-by: Chin Huang <[email protected]>
Fix issue 953 (#956) * fix onnx development version install issue in CI test Signed-off-by: degaochu <[email protected]> * fix windows issue Signed-off-by: degaochu <[email protected]> * fix windows issue Signed-off-by: degaochu <[email protected]> * fix windows issue Signed-off-by: degaochu <[email protected]> * fix windows issue Signed-off-by: degaochu <[email protected]> * fix windows issue Signed-off-by: degaochu <[email protected]> * exlucde unimplemented operator in onnx backend test Signed-off-by: degaochu <[email protected]> * remove .project Signed-off-by: degaochu <[email protected]> * fix issue 951 Signed-off-by: degaochu <[email protected]> * fix lgtm alert Signed-off-by: degaochu <[email protected]> * Update conv_mixin.py Signed-off-by: degaochu <[email protected]> * fix issue 953 Signed-off-by: degaochu <[email protected]> * Fix issue 953 Signed-off-by: degaochu <[email protected]>
Fix issue 951 (#954) * fix onnx development version install issue in CI test Signed-off-by: degaochu <[email protected]> * fix windows issue Signed-off-by: degaochu <[email protected]> * fix windows issue Signed-off-by: degaochu <[email protected]> * fix windows issue Signed-off-by: degaochu <[email protected]> * fix windows issue Signed-off-by: degaochu <[email protected]> * fix windows issue Signed-off-by: degaochu <[email protected]> * exlucde unimplemented operator in onnx backend test Signed-off-by: degaochu <[email protected]> * remove .project Signed-off-by: degaochu <[email protected]> * fix issue 951 Signed-off-by: degaochu <[email protected]> * fix lgtm alert Signed-off-by: degaochu <[email protected]> * Update conv_mixin.py Signed-off-by: degaochu <[email protected]>
Fix a bug in slice operator (#938) The len() method was used on a tensor, which is no longer legal in TF2. Signed-off-by: Michal Gallus <[email protected]> Co-authored-by: Chin Huang <[email protected]>
fix onnx development version install issue in CI test (#948) * fix onnx development version install issue in CI test Signed-off-by: degaochu <[email protected]> * fix windows issue Signed-off-by: degaochu <[email protected]> * fix windows issue Signed-off-by: degaochu <[email protected]> * fix windows issue Signed-off-by: degaochu <[email protected]> * fix windows issue Signed-off-by: degaochu <[email protected]> * fix windows issue Signed-off-by: degaochu <[email protected]> * exlucde unimplemented operator in onnx backend test Signed-off-by: degaochu <[email protected]> * remove .project Signed-off-by: degaochu <[email protected]>
Merge pull request #940 from chudegao/seqsupport support sequence type; support opset 14 of identity op
support generate opset for custom op/domain (#945) Signed-off-by: degaochu <[email protected]>
Add new operator Trilu (#936) * add new operator trilu Signed-off-by: degaochu <[email protected]> * update onnx version to master Signed-off-by: degaochu <[email protected]> * add dynamic shape support and test case; format code Signed-off-by: degaochu <[email protected]> * update input k to int64 Signed-off-by: degaochu <[email protected]>
Add HardSwish support (#932) Add HardSwish support, which is introduced in ONNX opset 14. Add also a filter to exclude new non-deterministic ONNX model tests. Signed-off-by: Chin Huang <[email protected]>
add opset 14 for relu (#934) Signed-off-by: degaochu <[email protected]>
Add allowzero support for reshape op (#930) Add allowzero support for reshape op which is introduced in opset 14. Signed-off-by: Chin Huang <[email protected]>
Fix Broken Link to More Tutorials (#927) Signed-off-by: Nikolai Janakiev <[email protected]> Co-authored-by: Chin Huang <[email protected]>
Add data types support for opset 14 (#926) Add data types support for opset 14 for the following ops - Add - Div - Mul - Sub Also fix an ONNX helper api name change in test_model.py Signed-off-by: Chin Huang <[email protected]>
Add support for Multinomial (#920) Add support for Multinomial as requested as part of issue https://github.com/onnx/onnx-tensorflow/issues/728 Signed-off-by: Chin Huang <[email protected]>
Fix unknown output shape for export (#918) Fix unknown output shape for export as reported in issue https://github.com/onnx/onnx-tensorflow/issues/913. Signed-off-by: Chin Huang <[email protected]>
Add version 1.8.0 to Versioning.md (#917) Add version 1.8.0 to Versioning.md Signed-off-by: Chin Huang <[email protected]>
Fix output names issue (#906) Fix output names issue https://github.com/onnx/onnx-tensorflow/issues/898 Also changed test code to process model outputs as dict. Signed-off-by: Chin Huang <[email protected]>