Integrate with ONNX 1.17.0 release branch

[roborags]

We are releasing ONNX 1.17.0. A release branch is created (https://github.com/onnx/onnx/tree/rel-1.17.0). The planned release date is September 30, 2024. Release candidates are also available from TestPyPI:

pip install -i https://test.pypi.org/simple/ --pre onnx

Note: If you system installs from the source tar.gz, you might encounter an error about the version of setuptools available when building onnx from the tar.gz. If so use:

pip install --upgrade -i https://test.pypi.org/simple/ --extra-index-url https://pypi.org/simple --pre onnx

It is important to integrate ONNX release branch into ORT ASAP so that any issues and incompatibilities can be detected and resolved before the ONNX release.

Key Updates

• ai.onnx Opset 22
  • Update to support bfloat16:
    • Acos, Acosh, Asin, Asinh, Atan, Atanh, AveragePool, Bernoulli, Conv, ConvTranspose, Cos, Cosh, DeformConv, Det, Dropout, Elu, EyeLike, GRU, GlobalAveragePool, GlobalLpPool, GlobalMaxPool, GridSample, HardSigmoid, HardSwish, InstanceNormalization, LSTM, LpNormalization, LpPool, MaxPool, MaxRoiPool, MaxUnpool, Mish, Multinomial, NegativeLogLikelihoodLoss, RNN, RandomNormal, RandomNormalLike, RandomUniform, RandomUniformLike, RoiAlign, Round, Selu, Sin, Sinh, Softplus, Softsign, Tan, ThresholdedRelu
• Support for numpy >= 2.0

In case a bug in ONNX is detected during integration of ONNX 1.17.0, please open a ONNX Bug Report and tag ONNX Release Manager - Raghavan @roborags so that the bug is fixed in the ONNX release branch.