Code

January 2025 Update#1036

Code

Merged

martindevans merged 25 commits intoSciSharp:masterSciSharp/LLamaSharp:masterfrom martindevans:wip_december_updatemartindevans/LLamaSharp:wip_december_updateCopy head branch name to clipboardJan 21, 2025

+563-237

Commits

Commits on Dec 3, 2024

• code changes for december update (not working yet)

  martindevans

  committed
  ac29c34

  Copy full SHA for ac29c34

Commits on Dec 20, 2024

• Changes to support up to ggerganov/llama.cpp@d408bb9

  martindevans

  committed
  e4f4fed

  Copy full SHA for e4f4fed

Commits on Dec 27, 2024

• Updated to latest llama.cpp binaries, this works on Windows CPU but needs more changes for other backends

  martindevans

  committed
  c90ddd9

  Copy full SHA for c90ddd9

Commits on Jan 4, 2025

• Updated to latest deps, fixed kernel memory failing to load

  martindevans

  committed
  c27cfde

  Copy full SHA for c27cfde

• Copy missing Mac flibraries libggml-base and libggml-cpu

  SignalRT

  committed
  a5c9759

  Copy full SHA for a5c9759

• Removed any mention of AVX in MacOS loading

  martindevans

  committed
  34198f9

  Copy full SHA for 34198f9

Commits on Jan 11, 2025

• Added file copying for some more targets (still missing macos)

  martindevans

  committed
  3d93174

  Copy full SHA for 3d93174

• Updated to latest set of binaries

  martindevans

  committed
  0647df9

  Copy full SHA for 0647df9

• Fixed copy path for CUDA12 DLLs

  martindevans

  committed
  756a88f

  Copy full SHA for 756a88f

Commits on Jan 17, 2025

• Compatibility with llama.cpp backend split (PR #10256) on all platforms

  m0nsky

  committed
  4950e0d

  Copy full SHA for 4950e0d

• Restore original comment

  m0nsky

  committed
  40a8c6c

  Copy full SHA for 40a8c6c

• Merge pull request #5 from m0nsky/wip_december_update_fixes

  Show description for 3521b27
  martindevans

  authored
  3521b27

  Copy full SHA for 3521b27

Commits on Jan 18, 2025

• Update the dependency loader for ggml-metal and ggml-blas

  m0nsky

  committed
  dc3dff1

  Copy full SHA for dc3dff1

• Update the runtime targets for ggml-metal and ggml-blas

  m0nsky

  committed
  7b558ce

  Copy full SHA for 7b558ce

• Add CPU backend (fallback) dependency for the GPU backends

  m0nsky

  committed
  6d0b421

  Copy full SHA for 6d0b421

• Fix icons for the nuget backends

  m0nsky

  committed
  4dbdc82

  Copy full SHA for 4dbdc82

Commits on Jan 19, 2025

• Update nuspec files for the GPU backends

  m0nsky

  committed
  556a7c1

  Copy full SHA for 556a7c1

• Update BinaryReleaseId

  m0nsky

  committed
  f526cbe

  Copy full SHA for f526cbe

• Update nuspec for CPU & OSX

  m0nsky

  committed
  91effe9

  Copy full SHA for 91effe9

• Merge pull request #6 from m0nsky/wip_december_update_fixes_v2

  Show description for 695a4da
  martindevans

  authored
  695a4da

  Copy full SHA for 695a4da

• Update CPU nuspec to use noavx folder

  m0nsky

  committed
  3be20b1

  Copy full SHA for 3be20b1

• Update Runtime.targets to use noavx folder

  m0nsky

  committed
  686627c

  Copy full SHA for 686627c

• Update BinaryReleaseId

  m0nsky

  committed
  1913966

  Copy full SHA for 1913966

Commits on Jan 20, 2025

• CUDA & Vulkan native libraries now correctly store the detected or user defined AVX level

  m0nsky

  committed
  014ef78

  Copy full SHA for 014ef78

• Merge pull request #7 from m0nsky/wip_december_update_fixes_v3

  Show description for 830a078
  martindevans

  authored
  830a078

  Copy full SHA for 830a078