* Check number of inputs, use rank function
* Change min inputs in flatten
* Fix code style
* Fix aten::tensor and aten::_convolution number of inputs
* Refactor NodeCOntext a little
* Fix codestyle
* [C++/BENCHMARK_APP] Introduce Numpy array loading for C++ benchmark app
* [DOCS/BENCHMARK_APP] Update docs to reflect changes, update list of available extensions from OpenCV, align help messages
* Update inputs_filling.cpp
* Update tools/benchmark_tool/openvino/tools/benchmark/utils/inputs_filling.py
Co-authored-by: Zlobin Vladimir <vladimir.zlobin@intel.com>
* Update samples/cpp/benchmark_app/inputs_filling.cpp
Co-authored-by: Zlobin Vladimir <vladimir.zlobin@intel.com>
* [C++/Python] Implement quality-of-life improvements from PR comments
* [C++] Fix compilation errors, fix linter output
* [C++/PYTHON] Apply requested changes
* Update samples/cpp/benchmark_app/main.cpp
Co-authored-by: Zlobin Vladimir <vladimir.zlobin@intel.com>
* Update samples/cpp/benchmark_app/utils.cpp
Co-authored-by: Zlobin Vladimir <vladimir.zlobin@intel.com>
* [PYTHON] Separate loading of numpy arrays similar to images
* [PYTHON] Remove unnecessary 'Prepare xxx file' print
* Update README again because IF OPENCV.. dissapeared for some reason
* Update second README with missing IF OPENCV..
* [C++] Remove unnecessary vector print function
* [C++ Add Numpy processing function - TODO link it to the tensor filling
* Reverse OneDnn plugin modification
* [C++] Numpy array loading for C++
* [C++] Add (almost) all missing types of data
* Reverse submodule modifications
* [C++/PYTHON] Fix compilation errors, clean code
* [C++] Modify supported extensions, add numpy checking to utils, add numpy to get_image_info method
* Update samples/cpp/benchmark_app/inputs_filling.cpp
Co-authored-by: Zlobin Vladimir <vladimir.zlobin@intel.com>
* [C++] Fix utils header file to reflect unordered set change
* [PYTHON/C++] Fix compilation errors in C++ code, fix Python dynamic shapes numpy loading
* [C++] Fix explicit instantiation of NumpyArray reader
* [C++] Clang format, minor syntax fixes
* [PYTHON/C++] Remove unnecessary data types, introduce a new approach to cast data of different types from format_rt_reader, remove uppercase types from Python precision parameters
* [PYTHON] Update README to reflect new precision settings
* [PYTHON] Fix README, fix clang format
* [C++] Clean headers
* [C++] Fix uninitialized variable error
* [C++/PYTHON] Fixed choices in Python benchmark, fixed types in C++ benchmark
* [C++] Fixed ov::float16conversion, fixed Python types map - removed redundancies
* [C++] Add back boolean support
* [C++] Fix compilation errors
---------
Co-authored-by: Zlobin Vladimir <vladimir.zlobin@intel.com>
* update test cases for windows cpu mapping parser
* fix code style issue
* fix code style issue
* fix code style issue
* update test data for Linux parser to align with Windows test case
* ufix code style issue
* update test data for Android build
* update linux CPU map parser fix
* fix code style issue
* fix test data issue
* fix size_t warning
* fix size_t warning
* serialization of proposal
* serialization of anchors in proposal
* added unit tests for gpu proposal
* updated the proposal primitive to be partially serialized
* serialization of primitive class
* removed unnecessary codes
* removed white spaces
* serialization of loop primitive
* serialization of nms
* fixed implicit concat logic in serialization
* added RUN_ALL_MODEL_CACHING_TESTS directive
* fixed an error related with specialization of boolean vector
* speed up constant folding and improved load time for FP16
* fix unit-tests
* added OPENVINO_SUPPRESS_DEPRECATED_START
* new evaluate for Convert; added constructor for Constant; constant copying by reference during constant_fold
* removed deprecation macro from Convert evaluate
* returned back Convert evaluate to old HostTensorPtr because it's used in tests accuracy comparator
* [TF FE] Refactor ZerosLike and add layer test
Signed-off-by: Kazantsev, Roman <roman.kazantsev@intel.com>
* Fix test for Wide and Deep model
---------
Signed-off-by: Kazantsev, Roman <roman.kazantsev@intel.com>
* [TF FE] Fix compilation warnings
Signed-off-by: Kazantsev, Roman <roman.kazantsev@intel.com>
* Fix compiling warnings for Einsum
* Apply code-review feedback
* Fix compilation warnings on Windows
* Only suppress warnings for MSVS
* Suppress false warning 4267 for extract_compressed_tensor_content
* Correct a place for end of warning suppress
---------
Signed-off-by: Kazantsev, Roman <roman.kazantsev@intel.com>
* Create output tensor for bound evaluate depends
on element type and shape, rank dynamism
* Add test to simulate pytorch bound evaluation
for when output has got dynamic data type or rank
* Correct code style
* Correct shape for dynamic tensor.
* Proposal of data_dispatcher
* Fix C-style for array-like objects
* Added test cases
* Add InferRequestInternal to hold lifetime of converted data
* Change name and location of InferRequestInternal to _InferRequestWrapper
* Fixes for shared data mode to match previous tests, codestyle refactoring
* Fix annotations and style
* Add initial testing
* Fix dtype capture
* Fix set_request_tensor
* Extend doc strings
* Fix docs
* Docs fix
* Change to ndarray in docs
* Fix async case
* Fix async test 2
* Remove linter issue
* [GPU] Added shape agnostic NonZero kernels
Signed-off-by: Andrew Park <andrew.park@intel.com>
* Add dynamic TC for ov_gpu_unit_tests
Signed-off-by: Andrew Park <andrew.park@intel.com>
* Apply code review
Signed-off-by: Andrew Park <andrew.park@intel.com>
---------
Signed-off-by: Andrew Park <andrew.park@intel.com>
