Merge branch 'main' into adrianl/BaseTester_SupportCompileAndKernelPluginEps

Author: adrianlizarraga

cmake/onnxruntime_unittests.cmake

@@ -1520,8 +1520,13 @@ endif()
                               onnxruntime_common ${CMAKE_DL_LIBS})
         set_target_properties(onnxruntime_runtime_path_test_shared_library PROPERTIES AIX_SHARED_LIBRARY_ARCHIVE OFF)
       else()
-        target_link_libraries(onnxruntime_runtime_path_test_shared_library PRIVATE
-                              onnxruntime_common cpuinfo ${CMAKE_DL_LIBS})
+        if (CPUINFO_SUPPORTED)
+          target_link_libraries(onnxruntime_runtime_path_test_shared_library PRIVATE
+                                onnxruntime_common cpuinfo ${CMAKE_DL_LIBS})
+        else()
+          target_link_libraries(onnxruntime_runtime_path_test_shared_library PRIVATE
+                                onnxruntime_common ${CMAKE_DL_LIBS})
+        endif()
       endif()
       target_include_directories(onnxruntime_runtime_path_test_shared_library PRIVATE ${ONNXRUNTIME_ROOT})
 

dockerfiles/Dockerfile.source

@@ -16,4 +16,4 @@ RUN cd /code && /bin/bash ./build.sh --allow_running_as_root --skip_submodule_sy
 FROM mcr.microsoft.com/azurelinux/base/python:3
 COPY --from=0 /code/build/Linux/Release/dist /root
 COPY --from=0 /code/dockerfiles/LICENSE-IMAGE.txt /code/LICENSE-IMAGE.txt
-RUN tdnf install -y ca-certificates python3-setuptools python3-wheel python3-pip python3-numpy python3-flatbuffers python3-packaging python3-protobuf python3-mpmath python3-sympy && python3 -m pip install coloredlogs humanfriendly && python3 -m pip install --no-index --find-links /root onnxruntime  && rm -rf /root/*.whl
+RUN tdnf install -y ca-certificates python3-setuptools python3-wheel python3-pip python3-numpy python3-flatbuffers python3-packaging python3-protobuf python3-mpmath python3-sympy && python3 -m pip install humanfriendly && python3 -m pip install --no-index --find-links /root onnxruntime  && rm -rf /root/*.whl

docs/OperatorKernels.md

@@ -240,7 +240,8 @@ Do not modify directly.*
 |||[13, 15]|**B** = tensor(bool)<br/> **I** = tensor(int64)<br/> **V** = seq(tensor(bfloat16)), seq(tensor(bool)), seq(tensor(double)), seq(tensor(float)), seq(tensor(float16)), seq(tensor(int16)), seq(tensor(int32)), seq(tensor(int64)), seq(tensor(int8)), seq(tensor(string)), seq(tensor(uint16)), seq(tensor(uint32)), seq(tensor(uint64)), seq(tensor(uint8)), tensor(bfloat16), tensor(bool), tensor(double), tensor(float), tensor(float16), tensor(int16), tensor(int32), tensor(int64), tensor(int8), tensor(string), tensor(uint16), tensor(uint32), tensor(uint64), tensor(uint8)|
 |||[11, 12]|**B** = tensor(bool)<br/> **I** = tensor(int64)<br/> **V** = tensor(bfloat16), tensor(bool), tensor(double), tensor(float), tensor(float16), tensor(int16), tensor(int32), tensor(int64), tensor(int8), tensor(string), tensor(uint16), tensor(uint32), tensor(uint64), tensor(uint8)|
 |||[1, 10]|**B** = tensor(bool)<br/> **I** = tensor(int64)<br/> **V** = tensor(bfloat16), tensor(bool), tensor(double), tensor(float), tensor(float16), tensor(int16), tensor(int32), tensor(int64), tensor(int8), tensor(string), tensor(uint16), tensor(uint32), tensor(uint64), tensor(uint8)|
-|LpNormalization|*in* input:**T**<br> *out* output:**T**|1+|**T** = tensor(double), tensor(float)|
+|LpNormalization|*in* input:**T**<br> *out* output:**T**|22+|**T** = tensor(double), tensor(float)|
+|||[1, 21]|**T** = tensor(double), tensor(float)|
 |LpPool|*in* X:**T**<br> *out* Y:**T**|22+|**T** = tensor(float)|
 |||[18, 21]|**T** = tensor(float)|
 |||[11, 17]|**T** = tensor(float)|

docs/python/requirements.txt

@@ -11,7 +11,6 @@ furo
 pyquickhelper
 pandas
 pydot
-coloredlogs
 flatbuffers
 numpy<2.0.0
 packaging

include/onnxruntime/core/framework/allocator.h

@@ -85,8 +85,8 @@ constexpr const char* OpenVINO_GPU = "OpenVINO_GPU";
 constexpr const char* OpenVINO_RT = "OpenVINO_RT";
 constexpr const char* OpenVINO_RT_NPU = "OpenVINO_RT_NPU";
 constexpr const char* QNN_HTP_SHARED = "QnnHtpShared";
-constexpr const char* WEBGPU_BUFFER = "WebGPU_Buffer";
-constexpr const char* WEBNN_TENSOR = "WebNN_Tensor";
+constexpr const char* WEBGPU_BUFFER = "WebGPU_Buf";  // limited to 10 chars to ensure std::string SSO for web
+constexpr const char* WEBNN_TENSOR = "WebNN_Ten";    // limited to 10 chars to ensure std::string SSO for web
 
 constexpr size_t kAllocAlignment = 256;
 constexpr const size_t kAlloc4KAlignment = 4096;

include/onnxruntime/core/session/onnxruntime_c_api.h

@@ -7196,11 +7196,36 @@ struct OrtApi {
    */
   ORT_API_T(void, RunOptionsSetSyncStream, _Inout_ OrtRunOptions* options, _In_ OrtSyncStream* sync_stream);
 
+  /** \brief Get the element data type and shape for an OrtValue that represents a Tensor (scalar, dense, or sparse).
+   *
+   * \note This function is an alternative to ::GetTensorTypeAndShape() that does not allocate a new array for
+   *       the shape data. The OrtValue instance's internal shape data is returned directly.
+   *
+   * \note Returns an error if the underlying OrtValue is not a Tensor.
+   *
+   * \param[in] value The OrtValue instance.
+   * \param[out] elem_type Output parameter set to the tensor element data type.
+   * \param[out] shape_data Output parameter set to the OrtValue instance's internal shape data array.
+   *                        For a scalar, `shape_data` is NULL and `shape_data_count` is 0.
+   *                        Must not be released as it is owned by the OrtValue instance. This pointer becomes invalid
+   *                        when the OrtValue is released or if the underlying shape data is updated or reallocated.
+   * \param[out] shape_data_count Output parameter set to the number of elements in `shape_data`.
+   *                              `shape_data_count` is 0 for a scalar.
+   *
+   * \snippet{doc} snippets.dox OrtStatus Return Value
+   *
+   * \since Version 1.24.
+   */
+  ORT_API2_STATUS(GetTensorElementTypeAndShapeDataReference, _In_ const OrtValue* value,
+                  _Out_ ONNXTensorElementDataType* elem_type,
+                  _Outptr_result_maybenull_ const int64_t** shape_data,
+                  _Out_ size_t* shape_data_count);
+
   /** \brief Enable profiling for this run
    *
    * \param[in] options
    * \param[in] profile_file_prefix The prefix for the profile file. The actual filename will be:
-   *                                <profile_file_prefix>_<timestamp>.json
+   *                                [profile_file_prefix]_[timestamp].json
    *
    * \snippet{doc} snippets.dox OrtStatus Return Value
    *

include/onnxruntime/core/session/onnxruntime_cxx_api.h

@@ -2233,6 +2233,19 @@ struct ConstValueImpl : Base<T> {
   const R* GetSparseTensorValues() const;
 
 #endif
+
+  /// <summary>
+  /// Returns the tensor's element type and a reference to the tensor's internal shape data. The shape data is owned
+  /// by the Ort::Value and becomes invalid when the Ort::Value is destroyed or if the underlying shape data is
+  /// updated or reallocated.
+  ///
+  /// For a scalar, shape.shape is nullptr and shape.shape_len is 0.
+  ///
+  /// Wraps OrtApi::GetTensorElementTypeAndShapeDataReference.
+  /// </summary>
+  /// <param name="elem_type">Output parameter set to the element's data type.</param>
+  /// <param name="shape">Output parameter set to the OrtValue instance's shape data and number of elements.</param>
+  void GetTensorElementTypeAndShapeDataReference(ONNXTensorElementDataType& elem_type, Shape& shape) const;
 };
 
 template <typename T>

include/onnxruntime/core/session/onnxruntime_cxx_inline.h

@@ -2387,6 +2387,13 @@ inline const R* ConstValueImpl<T>::GetSparseTensorValues() const {
 
 #endif
 
+template <typename T>
+void ConstValueImpl<T>::GetTensorElementTypeAndShapeDataReference(ONNXTensorElementDataType& elem_type,
+                                                                  Shape& shape) const {
+  ThrowOnError(GetApi().GetTensorElementTypeAndShapeDataReference(this->p_, &elem_type, &shape.shape,
+                                                                  &shape.shape_len));
+}
+
 template <typename T>
 void ValueImpl<T>::FillStringTensor(const char* const* s, size_t s_len) {
   ThrowOnError(GetApi().FillStringTensor(this->p_, s, s_len));