Xavier Dupré

[WIP] experimental: c export of ModelProto in python package

2

**Description** Many functions serialize a proto into bytes then deserialized in the C++ binding (checked, infershapes, ...) It could be faster by avoiding this step. This PR tries to make...

**Description** Many functions serialize a proto into bytes then deserialized in the C++ binding (checked, infershapes, ...) It could be faster by avoiding this step. This PR tries to make...

Update Versioning.md

3

**Description**: Update documentation. **Motivation and Context** - Versioning.md was not up to date.

Export schema in onnxruntime-training if --gen_doc is on

3

**Description**: Describe your changes. Function `get_all_operator_schema` is missing in the python binding of onnxruntime-training if flag --gen_doc is switched on. This PR add them. **Motivation and Context** This changes lets...

This error happens while running the unit test with the nightly build.

convert_keras fails with quantize_model

1

Following example fails with error ``AssertionError: dense_input:01 is disconnected, check the parsing log for more details.``. ``` model = quantize_model( keras.Sequential([ layers.Dense(3, activation='relu', input_shape=(5,)), layers.Dense(3, activation='relu', input_shape=(3,)), ])) model.compile(optimizer="Adam", loss="mse",...

[WIP] Fix conversion of XGBoost model after being restored

3

Investigate issue #499.

See https://github.com/microsoft/onnxruntime/issues/8008.