diff --git a/docs/articles_en/get_started/installing-openvino-overview.rst b/docs/articles_en/get_started/installing-openvino-overview.rst index 9f42f3de9a1..920d3a6ba9d 100644 --- a/docs/articles_en/get_started/installing-openvino-overview.rst +++ b/docs/articles_en/get_started/installing-openvino-overview.rst @@ -1,11 +1,11 @@ .. {#openvino_docs_install_guides_overview} -Install OpenVINO™ 2023.2 +Install OpenVINO™ 2023.3 ========================== .. meta:: - :description: install OpenVINO Runtime package, using the distribution channel + :description: install OpenVINO Runtime package, using the distribution channel of your choice. @@ -15,7 +15,7 @@ Install OpenVINO™ 2023.2 OpenVINO Runtime on Linux OpenVINO Runtime on Windows - OpenVINO Runtime on macOS + OpenVINO Runtime on macOS Create a Yocto Image @@ -26,8 +26,8 @@ Install OpenVINO™ 2023.2 .. warning:: - - The OpenVINO Development Tools package has been deprecated and removed from the default + + The OpenVINO™ Development Tools package has been deprecated and removed from the default installation options. For new projects, the OpenVINO runtime package now includes all necessary components. @@ -36,30 +36,30 @@ Install OpenVINO™ 2023.2 .. tip:: - - OpenVINO 2023.2, described here, is not a Long-Term-Support version! + + OpenVINO 2023.3, described here, is a Long-Term-Support version! All currently supported versions are: - * 2023.2 (development) + * 2023.3 (LTS) * 2022.3 (LTS) - * 2021.4 (LTS) + Moreover, different OpenVINO distributions may support slightly different sets of features. - Read installation guides for particular distributions for more details. + Read installation guides for particular distributions for more details. - .. dropdown:: Distribution Comparison for OpenVINO 2023.2 - - =============== ========== ====== ========= ======== ============ ========== ========== - Device Archives PyPI APT/YUM Conda Homebrew vcpkg Conan - =============== ========== ====== ========= ======== ============ ========== ========== - CPU V V V V V V V - GPU V V V V V V V - GNA V n/a n/a n/a n/a n/a n/a - NPU V n/a n/a n/a n/a n/a n/a - =============== ========== ====== ========= ======== ============ ========== ========== + .. dropdown:: Distribution Comparison for OpenVINO 2023.3 + + =============== ========== ====== ========= ======== ============ ========== ========== + Device Archives PyPI APT/YUM Conda Homebrew vcpkg Conan + =============== ========== ====== ========= ======== ============ ========== ========== + CPU V V V V V V V + GPU V V V V V V V + GNA V n/a n/a n/a n/a n/a n/a + NPU V n/a n/a n/a n/a n/a n/a + =============== ========== ====== ========= ======== ============ ========== ========== | **Build OpenVINO from source** -| OpenVINO Toolkit source files are available on GitHub as open source. If you want to build your own version of OpenVINO for your platform, +| OpenVINO Toolkit source files are available on GitHub as open source. If you want to build your own version of OpenVINO for your platform, follow the `OpenVINO Build Instructions `__. diff --git a/docs/articles_en/openvino_workflow/running_inference_with_openvino/Device_Plugins/NPU.rst b/docs/articles_en/openvino_workflow/running_inference_with_openvino/Device_Plugins/NPU.rst index 0b36f070930..f8e21a463b2 100644 --- a/docs/articles_en/openvino_workflow/running_inference_with_openvino/Device_Plugins/NPU.rst +++ b/docs/articles_en/openvino_workflow/running_inference_with_openvino/Device_Plugins/NPU.rst @@ -4,7 +4,7 @@ NPU Device ========== .. meta:: - :description: OpenVINO™ supports the Neural Processing Unit, + :description: OpenVINO™ supports the Neural Processing Unit, a low-power processing device dedicated to running AI inference. @@ -13,17 +13,21 @@ Intel® Core™ Ultra generation of CPUs (formerly known as Meteor Lake). It ena you to offload certain neural network computation tasks from other devices, for more streamlined resource management. +Note that the NPU plugin is currently available only with the Archive distribution of OpenVINO™ +and you need to :doc:`install a proper NPU driver ` +to use it successfully. For an in-depth description of the NPU plugin, see: + For an in-depth description of the NPU plugin, see: -• `NPU plugin developer documentation `__ -• `OpenVINO Runtime NPU plugin source files `__ +* `NPU plugin developer documentation `__ +* `OpenVINO Runtime NPU plugin source files `__ | **Supported Platforms:** | Host: Intel® Core™ Ultra (former Meteor Lake) | NPU device: NPU 3720 | OS: Ubuntu* 20, MS Windows* 11 (both 64-bit) - + | **Supported Inference Data Types** | The NPU plugin supports the following data types as inference precision of internal primitives: @@ -32,7 +36,7 @@ For an in-depth description of the NPU plugin, see: | Computation precision for the HW is fp16. | | For more details on how to get a quantized model, refer to the - :doc:`Model Optimization guide `, and + :doc:`Model Optimization guide `, and :doc:`NNCF tool quantization guide `. @@ -56,13 +60,13 @@ UMD Dynamic Model Caching +++++++++++++++++++++++++++++ UMD model caching is a solution enabled by default in the current NPU driver. -It improves time to first inference (FIL) by storing the model in the cache +It improves time to first inference (FIL) by storing the model in the cache after the compilation (included in FEIL), based on a hash key. The process may be summarized in three stages: 1. UMD generates the key from the input IR model and build arguments 2. UMD requests the DirectX Shader cache session to store the model - with the computed key. + with the computed key. 3. All subsequent requests to compile the same IR model with the same arguments use the pre-compiled model, reading it from the cache instead of recompiling. @@ -73,7 +77,7 @@ OpenVINO Model Caching OpenVINO Model Caching is a common mechanism for all OpenVINO device plugins and can be enabled by setting the ``ov::cache_dir`` property. This way, the UMD model caching is automatically bypassed by the NPU plugin, which means the model -will only be stored in the OpenVINO cache after compilation. When a cache hit +will only be stored in the OpenVINO cache after compilation. When a cache hit occurs for subsequent compilation requests, the plugin will import the model instead of recompiling it. @@ -86,8 +90,8 @@ Supported Features and properties The NPU device is currently supported by AUTO and MULTI inference modes. -The NPU support in OpenVINO is still under active development and may -offer a limited set of supported OpenVINO features. +The NPU support in OpenVINO is still under active development and may +offer a limited set of supported OpenVINO features. **Supported Properties:** @@ -110,11 +114,11 @@ offer a limited set of supported OpenVINO features. ov::intel_vpux::dpu_groups ov::intel_vpux::dma_engines ov::intel_vpux::compilation_mode - ov::intel_vpux::compilation_mode_params - ov::intel_vpux::print_profiling - ov::intel_vpux::profiling_output_file - ov::intel_vpux::vpux_platform - ov::intel_vpux::use_elf_compiler_backend + ov::intel_vpux::compilation_mode_params + ov::intel_vpux::print_profiling + ov::intel_vpux::profiling_output_file + ov::intel_vpux::vpux_platform + ov::intel_vpux::use_elf_compiler_backend .. tab-item:: Read-only properties @@ -132,7 +136,7 @@ offer a limited set of supported OpenVINO features. ov::intel_vpux::device_total_mem_size ov::intel_vpux::driver_version -.. note:: +.. note:: The optimum number of inference requests returned by the plugin based on the performance mode is **4 for THROUGHPUT** and **1 for LATENCY**. @@ -142,10 +146,10 @@ offer a limited set of supported OpenVINO features. Limitations ############################# -* Currently, only the models with static shapes are supported on NPU. +* Currently, only the models with static shapes are supported on NPU. * If the path to the model file includes non-Unicode symbols, such as in Chinese, - the model cannot be used for inference on NPU. It will return an error. -* Running the Alexnet model with NPU may result in a drop in accuracy. + the model cannot be used for inference on NPU. It will return an error. +* Running the Alexnet model with NPU may result in a drop in accuracy. At this moment, the googlenet-v4 model is recommended for classification tasks.