ngxson/llama.cpp

Pull Requests Commits

llama : DeepSeek V2/V3 MLA implementation (#12801)

jukofyork committed 254 days ago

Verified daa42288

ggml : Add AVX512 implementation of GEMM - Q4_Kx8 (#12829)

Srihari-mcw committed 254 days ago

Verified eccc7a16

CANN: Opt ROPE optimization (#12865)

noemotiovon committed 254 days ago

Verified 0019279b

CANN: Optimize CANN buffer pool memory management (#12875)

bachelor-dou committed 254 days ago

Verified b0c75ac9

Add performance print for gemma3 in example (#12929)

Russyyds committed 255 days ago

Verified d6d2c2ab

SYCL: Fix im2col (#12910)

qnixsynapse committed 255 days ago

Verified 75afa0ae

rpc : use ggml_context_ptr (#12938)

rgerganov committed 255 days ago

Verified c772d549

dsiable curl lib check, this action is missed by commit bd3f59f81289b920bcc597a208c14f55e39ed37e (#12761) (#12937)

NeoZhangJianyu committed 255 days ago

Verified 81c7e64f

ggerganov committed 255 days ago

526739b8

cpu: fix cpu backend's supports-op for GET_ROWS_BACK. fixes a fatal when running test-backend-ops with only the CPU backend (ggml/1190)

cmdr2 committed 255 days ago

a25355e2

ggml: use _mm[512/256]_dpbusd[_avx]_epi32 to directly accumulate into the result register (#12773)

SongXiaoXi committed 255 days ago

Verified e959d32b

ggml: disable CUDA graphs for unsupported DUP and CONT node types (#12891)

agray3 committed 256 days ago

Verified 307bfa25

quantize: Handle user-defined quantization levels for additional tensors (#12511)

EAddario committed 256 days ago

Verified 71e90e88

common : Define cache directory on AIX (#12915)

mehendarkarprajwal committed 257 days ago

Verified bc091a4d

vulkan: use aligned loads for flash attention mask (#12853)

jeffbolznv committed 257 days ago

Verified a4837577

llava: Fix cpu-only clip image encoding sefault (#12907)

mattjcly committed 257 days ago

Verified e59ea539

server : add VSCode's Github Copilot Chat support (#12896)

ggerganov committed 258 days ago

Verified c94085df

rpc : Set cache directory in rpc-server.cpp on FreeBSD (#12903)

yurivict committed 258 days ago

Verified e8a62631

`tool-call`: fix non-tool-calling grammar crashes w/ Qwen / Hermes 2 templates (#12900)

ochafik committed 258 days ago

Verified b6930ebc

common : Define cache directory on FreeBSD (#12892)

yurivict committed 258 days ago

Verified 68b08f36

sycl: Support sycl_ext_oneapi_limited_graph (#12873)

EwanC committed 258 days ago

Verified 578754b3

contrib: support modelscope community (#12664)

tastelikefeet committed 258 days ago

Verified b2034c2b

llama-model : add Glm4Model implementation for GLM-4-0414 (#12867)

zRzRzRzRzRzRzR committed 258 days ago

Verified 06bb53ad

clip : use smart pointer (⚠️ breaking change) (#12869)

ngxson committed 258 days ago

Verified 0c509239

SYCL: Add fp16 type support to unary op kernels (#12788)

qnixsynapse committed 258 days ago

Verified fccf9cae

convert : Llama4 RoPE fix (#12889)

danielhanchen committed 258 days ago

Verified ec6c09d0

ci : Replace freediskspace to free_disk_space in docker.yml (#12861)

yeahdongcn committed 258 days ago

Verified 8ac9f5d7

xcf : add check for visionos build version (#12854)

danbev committed 258 days ago

Verified 12e9158f

convert : proper tensor name mapping for llama4 (#12870)

ngxson committed 258 days ago

Verified 5b1f13cb

llama : correct rms norm for llama 4 (#12882)

ngxson committed 258 days ago

Verified 8b91d535

Newer Older