openvino
openvino copied to clipboard
OpenVINO™ is an open-source toolkit for optimizing and deploying AI inference
### Details: - Added variants of Philox algorithm generators from PyTorch and Tensorflow to align generated output for the same seed - Split execution into generation and conversion, reducing the...
### Details: - Improved LLMs latency for models with BF16 weights (by extending compressed IP kernel on BF16 data type) - Improved compilation time for models with BF16 weights on...
### Details: - Fixed tensors' shapes. ### Tickets: - CVS-137216
### Tickets: - *142049*
### Tickets: - *141573*
### OpenVINO Version openvino2022.3 ### Operating System Windows System ### Device used for inference CPU ### Framework PaddlePaddle ### Model used paddle OCR ### Issue description openvino version 22.3 can...
### Details: - by default clamping should be enabled same as for constant conversions ### Tickets: - CVS-141762
### Details: - By this PR we want to create `ov_npu_func_tests` executable having functional tests from `NPU plugin` (SLTs and Subgraphs to be added later). - Functional tests were copy-pasted...
### Details: - Windows std::wstring convert to std::string some unicode will be ???, such as L"unicode_Яㅎあ" -> "unicode_???" - Solution using the same way as Linux ### Tickets: - CVS-141444...
### Details: - *I lost code `t_stream_count_map[(void*)this] = item.first;` in #19832* - *the thread safe issue happen in below workflow* - create thread A call CustomThreadLocal:local() in thread A ->...