Skip to content

JavaCPP Presets 1.5.13

Latest
Latest

Choose a tag to compare

View all tags
@saudet saudet released this 22 Feb 05:12
· 14 commits to master since this release
1.5.13
This tag was signed with the committer’s verified signature.
saudet Samuel Audet
GPG key ID: EEC319672E4D0671
Verified
Learn about vigilant mode.
60df69b
This commit was signed with the committer’s verified signature.
saudet Samuel Audet
GPG key ID: EEC319672E4D0671
Verified
Learn about vigilant mode.

February 22, 2026 version 1.5.13

  • Ensure ONNX Runtime libraries get linked with -z noexecstack on Linux (issue #1715)
  • Add setter methods for GELUOptions.approximate() in presets for PyTorch (pull #1733)
  • Build FFmpeg with HarfBuzz to reenable the drawtext filter (pull #1731)
  • Enable training APIs and include onnxruntime_training_c_api.h header file in presets for ONNX Runtime (pull #1728)
  • Fix memory leak in Tensor factory methods create(), createBuffer(), and createIndexer() for PyTorch (pull #1722)
  • Rename gloo and nccl modules of PyTorch to torch_gloo and torch_nccl to avoid conflicts
  • Upgrade to NDK r29 for Android builds on GitHub Actions (issue #1670)
  • Enable distributed package using NCCL in presets for PyTorch (issue #1402)
  • Fix adapter for c10::intrusive_ptr in presets for PyTorch (issue #1678)
  • Include autocast_mode.h, FileStore.hpp, TCPStore.hpp, and gloo/transport/tcp/device.h header files in presets for PyTorch (discussion #1712)
  • Introduce linux-arm64 builds for ONNX (pull #1697)
  • Include CUDA GPU compute capabilities 10.0 and 12.0 in builds for PyTorch (issue #1696)
  • Enable DNNL and CoreML in linux-arm64 and macosx-arm64 builds for ONNX Runtime (pull #1689)
  • Refactor the redist artifacts for CUDA into multiple smaller main artifacts (issue #1638)
  • Introduce macosx-arm64 builds for DNNL, OpenCL (issue #1681)
  • Upgrade presets for system APIs on Ubuntu 22.04 LTS (Jammy), macOS 15, Windows SDK 10.0 (pull #1659)
  • Add windows-arm64 builds for LLVM (pull #1656)
  • Introduce CUDA-enabled linux-arm64-gpu builds to presets for OpenCV (pull #1651)
  • Reenable again linux-arm64 builds for CPython and NumPy (pull #1652)
  • Bundle libraries for VA-API 2.x with FFmpeg to avoid loading issues (issue bytedeco/javacv#2340)
  • Upgrade presets for OpenCV 4.13.0, FFmpeg 8.0.1, MKL 2025.3.1, DNNL 3.11, OpenBLAS 0.3.31, CUDA 13.1.1, cuDNN 9.19.0.56, NCCL 2.29.3, nvCOMP 5.1.0.21, NVIDIA Video Codec SDK 13.0.37, CPython 3.14.3, NumPy 2.4.2, SciPy 1.17.0, LLVM 21.1.8, libffi 3.5.2, Leptonica 1.87.0, Tesseract 5.5.2, PyTorch 2.10.0, SentencePiece 0.2.1, TensorFlow Lite 2.20.0, TensorRT 10.15.1.29, Triton Inference Server 2.65.0, ONNX 1.20.1, ONNX Runtime 1.24.1, and their dependencies
Assets 4
Loading