GPT2_LM_HEAD is a new ONNX model zoo model that OpenVino doesn't support.
Error message:1: [ONNXRuntimeError] : 6 : RUNTIME_EXCEPTION : Non-zero status code returned while running OpenVINO-EP-subgraph_1162 node. Name:'OpenVINOExecutionProvider_OpenVINO-EP-subgraph_1162_1' Status Message: _Map_base::at
The bug happens when merging following shapes:
input0: [1, 1, 'Min(1024, input1_dynamic_axes_3)', 'Min(1024, input1_dynamic_axes_3)']
input1: ['input1_dynamic_axes_1*input1_dynamic_axes_2', 12, 'input1_dynamic_axes_3', 'input1_dynamic_axes_3']
input2: []
The fix is to avoid broadcasting merge on input2
- Update docker image release build to use build commit.
- Use valid default in component governance detection step.
- Use smaller docker build context.
* Fix Windows AI version
* Update text to extend telemetry coverage
Includes all official binaries
* Update text about EP pluggability
* Update CUDA/cuDNN versions
* Add link to reduce operator kernel page
* Update roadmap
* Add preview for migraphx
* Move Rockchip under IoT/Edge
* Update text to include ORT for Mobile doc link
Add additional test cases such as ThreadPoolTest.TestPoolCreation_10Iter to create and destroy thread pools to watch for any memory leaks. Running under Valgrind, these tests should show all of the data allocated being deallocated again.
Two recent issues #5176 and #5292 indicated memory leaks. The test cases help identify whether or not any of the data structures used in the thread pool are being leaked. Currently, on WSL, the only data not being de-allocated in these tests are a small number of nsync waiter objects. This behavior is as expected (the waiter objects should be held on a free list in the nsync library).
* Add iOS test pipeline and a sample app.
* clean up the unused code.
* clean up.
* revert the unknown change
* disable the shared library for iOS.
* add open source notice text.
* ignore the skipped test.
* extract the common ortenv setup
* Move fbs include from header to cc
* add initial cmake for flatbuffers
* Move most flatbuffers util to ort_flatbuffers
* move code around
* fix
* move test/perf runner to use flatbuffer directly instead of model
* minor update
* Fix build break
* Clean up includes and foward decl
* Fix traning CI build breaks
* Addressed PR comment, replaced some include with forward decls
* Remove ORT_MUST_USE_RESULT temporarily
* Build Recomputation Graph
* Make topological sort to run FW nodes first
* Pattern match start and end of transformer layer
* Topological sort with Priority
* Add logger to Gradient Graph Builder
* Use Logger
* Introduce Execution Order
The definitions for some Eigen classes don't get pulled in leading to errors. Split out the broadcast function creation logic from the functions using std::enable_if to workaround that.
* Symbolic shape inference: fix a case when concat requires merge multiple dims
* Fix a bug triggered in newer version of sympy
Fix a bug in output data type guessing
* add custom logger and global threadpools to C and C++ API
* code cleanup and formatting
* reformat code
* tidy up some more code formatting
* remove comment
* fix API break from merging from master
* renamed API function to CreateEnvWithCustomLoggerAndGlobalThreadPools
* rename log variable and apply clang-format
* bug fix transformer
* fuse cpu kernel for transposescalematmul and matmul
* fuse transpose_scale_matmul cpu kernel with matmul
* fix test
* Add FusedMatMul Contrib Op
* fix test
* fix typo
* plus more updates per review
