Skip to content

Releases: ggml-org/llama.cpp

b5028

02 Apr 14:12
e0e912f
Compare
Choose a tag to compare
llama : add option to override model tensor buffers (#11397)

* llama : add option to override tensor buffers

* ggml : fix possible underflow in ggml_nbytes

b5026

02 Apr 11:28
83a88bd
Compare
Choose a tag to compare
vocab : BailingMoE : change possessive quantifiers to greedy (#12677)

b5025

02 Apr 10:50
42eb248
Compare
Choose a tag to compare
common : remove json.hpp from common.cpp (#12697)

* common : remove json.hpp from common.cpp

* fix comment

b5022

01 Apr 17:41
f423981
Compare
Choose a tag to compare
opencl : fix memory allocation size (#12649)

issue:
https://github.com/CodeLinaro/llama.cpp/pull/17#issuecomment-2760611283

This patch fixes the memory allocation size
not exceeding the maximum size of the OpenCL device.

b5021

01 Apr 13:42
e39e727
Compare
Choose a tag to compare
llama : use LLM_KV_GENERAL_FILE_TYPE instead of gguf_find_key (#12672)

b5019

01 Apr 12:40
3fd072a
Compare
Choose a tag to compare
metal : use F32 prec in FA kernels (#12688)

* metal : use F32 prec in FA kernels

ggml-ci

* cont : fix FA vec kernel

ggml-ci

b5018

01 Apr 12:10
a6f32f0
Compare
Choose a tag to compare
Fix clang warning in gguf_check_reserved_keys (#12686)

* Fix clang warning in gguf_check_reserved_keys

Signed-off-by: Xiaodong Ye <[email protected]>

* Fix typo

Signed-off-by: Xiaodong Ye <[email protected]>

---------

Signed-off-by: Xiaodong Ye <[email protected]>

b5017

01 Apr 10:22
2bb3597
Compare
Choose a tag to compare
vulkan: fix build when glslc doesn't support coopmat (#12683)

b5016

01 Apr 09:33
8293970
Compare
Choose a tag to compare
SYCL: Rename oneMKL to oneMath (#12192)

* Rename oneMKL Interface to oneMath

* Use oneMath for Intel vendor

* Rename occurences to mkl

* clang-format

* Silence verbose warnings

* Set oneMath HIP_TARGETS

* Fix silence warnings

* Remove step to build oneMath from build instructions

* Use fixed oneMath version

* Remove INTEL_CPU

* Fold CMake oneDNN conditions

* Use Intel oneMKL for Intel devices

* Improve CMake message

* Link against MKL::MKL_SYCL::BLAS only

* Move oneMath documentation to Nvidia and AMD sections

b5015

01 Apr 09:16
8bbf260
Compare
Choose a tag to compare
SYCL: switch to SYCL namespace (#12674)