diff --git a/docs/resources/prerelease_information.md b/docs/resources/prerelease_information.md index 162d4e16c85..527b9e7f737 100644 --- a/docs/resources/prerelease_information.md +++ b/docs/resources/prerelease_information.md @@ -30,10 +30,92 @@ Please file a github Issue on these with the label “pre-release” so we can g :material-regular:`feedback;1.4em` Share your feedback -.. dropdown:: OpenVINO Toolkit 2023.1.0.dev20230728 + + +.. dropdown:: OpenVINO Toolkit 2023.1.0.dev20230811 :animate: fade-in-slide-down :color: primary :open: + + `Check on GitHub `__ + + **New features:** + + * CPU runtime: + + * Enabled weights decompression support for Large Language models (LLMs). The implementation + supports avx2 and avx512 HW targets for Intel® Core™ processors and gives up to 2x improvement + in the latency mode (FP32 VS FP32+INT8 weights comparison). For 4th Generation Intel® Xeon® + Scalable Processors (formerly Sapphire Rapids) this INT8 decompression feature gives 10-25% + performance improvement, compared to pure BF16 inference. + * Reduced memory consumption of compile model stage by moving constant folding of Transpose + nodes to the CPU Runtime side. + * Set FP16 inference precision by default for non-convolution networks on ARM. Convolution + network will be executed in FP32. + + * GPU runtime: Added paddings for dynamic convolutions to improve performance for models like + Stable-Diffusion v2.1. + + * Python API: + + * Added the ``torchvision.transforms`` object to OpenVINO preprocessing. + * Moved all python tools related to OpenVINO into a single namespace, + improving user experience with better API readability. + + * TensorFlow FE: + + * Added support for the TensorFlow 1 Checkpoint format. All native TensorFlow formats are now enabled. + * Added support for 8 new operations: + + * MaxPoolWithArgmax + * UnravelIndex + * AdjustContrastv2 + * InvertPermutation + * CheckNumerics + * DivNoNan + * EnsureShape + * ShapeN + + * PyTorch FE: + + * Added support for 6 new operations. To know how to enjoy PyTorch models conversion follow + this `Link `__ + + * aten::concat + * aten::masked_scatter + * aten::linspace + * aten::view_as + * aten::std + * aten::outer + * aten::broadcast_to + + **New openvino_notebooks:** + + * `245-typo-detector `__ + : English Typo Detection in sentences with OpenVINO™ + + * `247-code-language-id `__ + : Identify the programming language used in an arbitrary code snippet + + * `121-convert-to-openvino `__ + : Learn OpenVINO model conversion API + + * `244-named-entity-recognition `__ + : Named entity recognition with OpenVINO™ + + * `246-depth-estimation-videpth `__ + : Monocular Visual-Inertial Depth Estimation with OpenVINO™ + + * `248-stable-diffusion-xl `__ + : Image generation with Stable Diffusion XL + + * `249-oneformer-segmentation `__ + : Universal segmentation with OneFormer + + +.. dropdown:: OpenVINO Toolkit 2023.1.0.dev20230728 + :animate: fade-in-slide-down + :color: secondary `Check on GitHub `__ @@ -55,7 +137,7 @@ Please file a github Issue on these with the label “pre-release” so we can g * Frameworks: - - PyTorch Updates: OpenVINO now supports originally quantized PyTorch models, including models produced with the Neural Network Compression Framework (NNCF). + - PyTorch Updates: OpenVINO now supports originally quantized PyTorch models, including models produced with the Neural Network Compression Framework (NNCF). - TensorFlow FE: Now supports Switch/Merge operations, bringing TensorFlow 1.x control flow support closer to full compatibility and enabling more models. - Python API: Python Conversion API is now the primary conversion path, making it easier for Python developers to work with OpenVINO.