ngxson/llama.cpp

Pull Requests Commits
cpu: de-duplicate some of the operators and refactor (ggml/1144)
cmdr2 committed 1 year ago
a62d7fa7
ggml : add logging for native build options/vars (whisper/2935)
danbev committed 1 year ago
e408d435
examples : command.wasm updates (whisper/2904)
danbev committed 1 year ago
3891e183
llama : fix non-causal mask for gemma 3 (#12615)
ngxson committed 1 year ago
Verified af6ae1ef
llama : change cpu_buft_list order: ACCEL -> GPU host -> CPU extra -> CPU (#12632)
Djip007 committed 1 year ago
Verified 0bb29193
cmake : fix ccache conflict (#12522)
BusyJay committed 1 year ago
Verified a69f8463
CANN : remove clang-format in ggml-cann (#12607)
hipudding committed 1 year ago
Verified d07a0d7a
llama : fix incorrect Qwen2Moe ffn_moe_out graph callback (#12631)
CISC committed 1 year ago
Verified 3714c3ee
metal : improve FA + improve MoE (#12612)
ggerganov committed 1 year ago
Verified b4ae5081
vulkan: fix coopmat shader generation when cross-compiling (#12272)
Icenowy committed 1 year ago
Verified b86f6007
llama: fix error on bad grammar (#12628)
JohannesGaessler committed 1 year ago
Verified dd373dd3
server : include speculative decoding stats when timings_per_token is enabled (#12603)
mostlygeek committed 1 year ago
Verified 5d016702
rpc : update README for cache usage (#12620)
rgerganov committed 1 year ago
Verified ef03229f
llamafile : ppc64le GEMV forwarding for FP32. (#12594)
amritahs-ibm committed 1 year ago
Verified 13731766
rpc : send hash when tensor data is above some fixed threshold (#12496)
rgerganov committed 1 year ago
Verified ab6ab8f8
server : Support listening on a unix socket (#12613)
Piotr committed 1 year ago
Verified 2099a9d5
media : add SVG logo [no ci] (#12616)
ggerganov committed 1 year ago
Verified 29690198
opencl: add multi and vision rope, `gelu_quick` and `im2col` (#12600)
lhez committed 1 year ago
Verified 5dec47dc
llama : add PLM GGUF Conversion & Inference Support (#12457)
Si1w committed 1 year ago
Verified f125b8dc
model : restore support for T5Encoder (#12590)
HighDoping committed 1 year ago
Verified 953c2a62
convert : Support Qwen2_5_VLForConditionalGeneration (#12595)
csabakecskemeti committed 1 year ago
Verified d5c6309d
sync : ggml
ggerganov committed 1 year ago
029c693f
scripts : update sync + fix cmake merge
ggerganov committed 1 year ago
771d8437
sync : ggml
ggerganov committed 1 year ago
df0665a4
cmake : sync/merge PowerPC build commands (#0)
ggerganov committed 1 year ago
0306aad1
llamafile : ppc64le MMA implementation for Q4_0. (#12489)
amritahs-ibm committed 1 year ago
Verified c7b43ab6
ggml : riscv: add 128-bit RVV support (#12530)
xctan committed 1 year ago
Verified 24feaec0
llama : make loras compatible with repacking (#12593)
ggerganov committed 1 year ago
Verified f28bc4c2
SYCL: implement memset ggml backend buffer interface (#12580)
qnixsynapse committed 1 year ago
Verified f17a3bb4
HIP: Add support for RDNA4 targets (#12372)
slojosic-amd committed 1 year ago
Verified bd40678d
Terms Of Use Privacy Policy Refund Policy Impressum