ochafik/llama.cpp

Pull Requests Commits
ggml : faster ssm scan (#10558)
A3shTnT committed 1 year ago
Verified 250d7953
convert : Qwerky : use lora_rank_tokenshift and lora_rank_decay if present (#12667)
CISC committed 1 year ago
Verified 403fbacb
Vulkan: Add DP4A MMQ and Q8_1 quantization shader (#12135)
0cc4m committed 1 year ago
Verified a8a1f335
cmake : fix whitespace (#0)
ggerganov committed 1 year ago
1790e731
sync : ggml
ggerganov committed 1 year ago
0114a32d
cmake: improve Vulkan cooperative matrix support checks (whisper/2966)
sandrohanea committed 1 year ago
a7724480
llava : proper description fix (#12668)
CISC committed 1 year ago
Verified 1a859490
SYCL: Remove misleading ggml_sycl_op_flatten function (#12387)
qnixsynapse committed 1 year ago
Verified 6c02a032
llava : fix clip loading GGUFs with missing description (#12660)
CISC committed 1 year ago
Verified f52d59d7
tts : remove printfs (#12640)
marcoStocchi committed 1 year ago
Verified 52de2e59
llama : support BailingMoE (Ling) (#12634)
CISC committed 1 year ago
Verified 2c3f8b85
metal : use constexpr in FA kernels + fix typedef (#12659)
ggerganov committed 1 year ago
Verified 4663bd35
llama : add Trillion 7B model support (#12556)
juyoung-trl committed 1 year ago
Verified b3de7cac
llama-chat : Add Yandex instruct model template support (#12621)
vorobyov01 committed 1 year ago
Verified 7242dd96
musa: fix all warnings, re-enable `-DLLAMA_FATAL_WARNINGS=ON` in ci and update doc (#12611)
yeahdongcn committed 1 year ago
Verified 492d7f1f
sync : ggml
ggerganov committed 1 year ago
d3f1f0ac
cpu : rm unused variable (ggml/1166)
ngxson committed 1 year ago
360dc22c
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
Terms Of Use Privacy Policy Refund Policy Impressum