Skip to content

JavaCPP Presets 1.5.8
Compare
Choose a tag to compare
@saudet saudet released this 02 Nov 14:38
· 146 commits to master since this release
1.5.8
This tag was signed with the committer’s verified signature.
saudet Samuel Audet
GPG key ID: EEC319672E4D0671
Learn about vigilant mode.
9934c22
This commit was signed with the committer’s verified signature.
saudet Samuel Audet
GPG key ID: EEC319672E4D0671
Learn about vigilant mode.

November 2, 2022 version 1.5.8

  • Fix mapping of torch::ExpandingArrayWithOptionalElem in presets for PyTorch (issue #1250)
  • Disable OpenMP for Tesseract to work around performance issues (pull #1201)
  • Enable NVIDIA GPU acceleration for FFmpeg on ARM and POWER as well (issue bytedeco/javacv#1894)
  • Add support for HEVC and Opus to FLV format in presets for FFmpeg (pull #1228)
  • Build Leptonica with OpenJPEG for JPEG 2000 support
  • Introduce linux-arm64 and macosx-arm64 builds to presets for libpostal (pull #1199 and pull #1205)
  • Map missing factory functions in torch:: namespace using torch_ as prefix in presets for PyTorch (issue #1197)
  • Add presets for the nvJPEG module of CUDA (issue #1193)
  • Introduce Android builds for TensorFlow Lite (discussion #1180)
  • Map std::vector<cv::Ptr<cv::mcc::CChecker> > for CCheckerDetector.getListColorChecker() (issue bytedeco/javacpp#571)
  • Include missing opencv2/mcc/ccm.hpp header file in presets for OpenCV (discussion bytedeco/javacpp#568)
  • Fix a few incorrectly mapped instances of std::unordered_map for PyTorch (issue #1164)
  • Migrate builds for Leptonica and Tesseract to CMake (issue #1163)
  • Introduce macosx-arm64 builds for LZ4 (pull #1243), libffi (issue #1182), Leptonica, and Tesseract (issue #814)
  • Map instances of torch::OrderedDict using C++ templates from PyTorch (issue #623)
  • Add presets for Bullet Physics SDK 3.24 (pull #1153)
  • Add long[] pytorch.Tensor.shape() method for convenience (pull #1161)
  • Enable DNNL codegen as BYOC backend in presets for TVM
  • Allow passing raw pointer as deleter to from_blob(), etc functions of PyTorch (discussion #1160)
  • Include cudnn_backend.h header file in presets for CUDA (issue #1158)
  • Bundle zlibwapi.dll required by cuDNN in redist artifacts of presets for CUDA (issue bytedeco/javacv#1767)
  • Harmonize string and buffer pointer types of function parameters from DepthAI (issue #1155)
  • Bundle correctly OpenMP library for PyTorch builds on Mac as well (issue #1225)
  • Remove dependency on CUDA from presets for Triton Inference Server (pull #1151)
  • Disable signal handlers of DepthAI known to cause issues with the JDK (issue #1118)
  • Upgrade presets for OpenCV 4.6.0, FFmpeg 5.1.2, HDF5 1.12.2, LZ4 1.9.4, MKL 2022.2, DNNL 2.7.1, OpenBLAS 0.3.21 (issue #1171), CPython 3.10.8, NumPy 1.23.4, SciPy 1.9.3, Gym 0.26.2, LLVM 15.0.3, libffi 3.4.4, Tesseract 5.2.0, CUDA 11.8.0, cuDNN 8.6.0, NCCL 2.15.5, OpenCL 3.0.12, MXNet 1.9.1, PyTorch 1.12.1, TensorFlow Lite 2.10.0, TensorRT 8.4.3.1, Triton Inference Server 2.26.0, ALE 0.8.0, DepthAI 2.18.0, ONNX 1.12.0, ONNX Runtime 1.13.1, TVM 0.10.0, Skia 2.88.3, cpu_features 0.7.0, ModSecurity 3.0.8, and their dependencies
Assets 6
Loading