[!NOTE] This patch release includes important bug fixes to the function definition of Attention-23/24 under the Group Query Attention mode and to the reference implementation of RotaryEmbedding-23.

All changes

• Avoid unnecessary re-generating of proto files (#7253) in https://github.com/onnx/onnx/pull/7306
• Require ml_dtypes>=0.5.0 (#7254) in https://github.com/onnx/onnx/pull/7307
• Cherry pick four attention PRs in https://github.com/onnx/onnx/pull/7315
• Update rotary_embedding reference implementation and tests (#7304, #7316) in https://github.com/onnx/onnx/pull/7313
• Override __repr__ for some proto classes (#7259) in https://github.com/onnx/onnx/pull/7314
• add check for rc-candidates (Update create_release.yml) (#7261) in https://github.com/onnx/onnx/pull/7323
• Implement repr methods for Model/Graph/Function (#7320) in https://github.com/onnx/onnx/pull/7325