ngxson/llama.cpp

Pull Requests Commits

CANN: Add support for async operator submission (#12864)

hipudding committed 296 days ago

Verified 7a395f67

llama : recognize IBM Granite 3.3 FIM tokens (#12988)

Noeda committed 296 days ago

Verified 971f245b

opencl: fix incorrect local_size index in profiling log (#12868)

kimminsu38oo committed 296 days ago

Verified 12b17501

vulkan: enable coopmat2 FA gqa and split_k optimizations more often (#12931)

jeffbolznv committed 296 days ago

Verified 015022bb

CANN: Add 310P operator support check (#12962)

noemotiovon committed 297 days ago

Verified b43d89e3

opencl: split `ggml-opencl.cl` into multiple files and cleanup (#12886)

lhez committed 297 days ago

Verified 80f19b41

metal : add FA-vec kernels for head size 96 (#12952)

ggerganov committed 298 days ago

Verified f8f820cc

CANN: Add x86 build ci (#12950)

hipudding committed 298 days ago

Verified 54a72720

CUDA/HIP: Share the same unified memory allocation logic. (#12934)

hjc4869 committed 298 days ago

Verified 84778e97

SYCL: Add ROPE vision kernel (#12887)

qnixsynapse committed 298 days ago

Verified 51067647

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

jukofyork committed 298 days ago

Verified daa42288

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

Srihari-mcw committed 298 days ago

Verified eccc7a16

CANN: Opt ROPE optimization (#12865)

noemotiovon committed 298 days ago

Verified 0019279b

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

bachelor-dou committed 298 days ago

Verified b0c75ac9

Add performance print for gemma3 in example (#12929)

Russyyds committed 298 days ago

Verified d6d2c2ab

SYCL: Fix im2col (#12910)

qnixsynapse committed 299 days ago

Verified 75afa0ae

rpc : use ggml_context_ptr (#12938)

rgerganov committed 299 days ago

Verified c772d549

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

NeoZhangJianyu committed 299 days ago

Verified 81c7e64f

ggerganov committed 299 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 299 days ago

a25355e2

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

SongXiaoXi committed 299 days ago

Verified e959d32b

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

agray3 committed 299 days ago

Verified 307bfa25

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

EAddario committed 299 days ago

Verified 71e90e88

common : Define cache directory on AIX (#12915)

mehendarkarprajwal committed 300 days ago

Verified bc091a4d

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

jeffbolznv committed 301 days ago

Verified a4837577

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

mattjcly committed 301 days ago

Verified e59ea539

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

ggerganov committed 301 days ago

Verified c94085df

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

yurivict committed 301 days ago

Verified e8a62631

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

ochafik committed 301 days ago

Verified b6930ebc

common : Define cache directory on FreeBSD (#12892)

yurivict committed 301 days ago

Verified 68b08f36

Newer Older