ngxson/llama.cpp

Pull Requests Commits

Update tools/mtmd/mtmd-cli.cpp

ngxson committed 236 days ago

Verified 16786523

mtmd-cli : fix mem leak

ngxson committed 236 days ago

6aaefb7c

mtmd : fix memory in mtmd_helper_eval_chunk_single

ngxson committed 237 days ago

a158e0fa

gguf: fix failure on version == 0 (#13956)

JohannesGaessler committed 237 days ago

Verified 7675c555

convert : fix nomic-bert-moe mask token (#13757)

CISC committed 237 days ago

Verified 5e1c3aed

convert : fix vocab padding code for bert models (#13954)

CISC committed 237 days ago

Verified c496fe0b

ggml: check if non-native endian model is being loaded (#13943)

taronaeo committed 237 days ago

Verified e57bb87c

ggerganov committed 237 days ago

f3a4b165

vulkan : Remove unexpected ; (ggml/1253)

dg0yt committed 237 days ago

108009f5

cmake : Fix broken CMake error messages (ggml/1252)

dg0yt committed 237 days ago

d337252a

ggml : remove ggml_graph_import and ggml_graph_export declarations (ggml/1247)

rgerganov committed 237 days ago

af6f91db

sync : whisper.cpp (ggml/1250)

ggerganov committed 237 days ago

a7b8d35f

ggml : install dynamic backends (ggml/1240)

rgerganov committed 237 days ago

6eba72b7

ggml : Print backtrace on uncaught C++ exceptions (ggml/1232)

danielzgtg committed 237 days ago

fedf034a

readme : update bindings (#13950)

ddh0 committed 237 days ago

Verified 8726392d

parallel : fix n_junk == 0 (#13952)

ggerganov committed 237 days ago

Verified c0462171

kv-cache : split implementation in separate sources (#13920)

ggerganov committed 237 days ago

Verified 0fc16b42

threading: support for GGML_SCHED_PRIO_LOW, update thread info on Windows to avoid throttling (#12995)

max-krasnyansky committed 238 days ago

Verified 053b1539

docs : Note about necessity of having libcurl installed for standard build. (#13945)

jpodivin committed 238 days ago

Verified b3a89c3d

server: allow unclosed thinking tags (#13931)

ochafik committed 238 days ago

Verified e15898d1

llama : deprecate explicit kv_self defrag/update calls (#13921)

ggerganov committed 238 days ago

Verified 803f8baf

llama : use n_swa + n_ubatch cells for SWA cache (#13833)

ggerganov committed 238 days ago

Verified 3600cc28

webui : Replace alert and confirm with custom modals. (#13711)

igardev committed 238 days ago

Verified c7e0a205

llama : auto-batch preparation (#13845)

ggerganov committed 238 days ago

Verified 3f55f781

mtmd : drop `_shared` from `libmtmd` name, merge helpers into libmtmd (⚠️ breaking change) (#13917)

ngxson committed 238 days ago

Verified 51fa76f1

kv-cache : refactor + add llama_memory_state_i (#13746)

ggerganov committed 238 days ago

Verified 12d0188c

CUDA: add a prop in ggml_cuda_device_infor for distinguish iGPU or dGPU in cuda (#13856) (#13895)

Yangxiaoz committed 238 days ago

Verified eb394993

CUDA: fix typo in FlashAttention code (#13926)

JohannesGaessler committed 239 days ago

Verified e562eece

sched : avoid changing cur_copy when a graph is already allocated (#13922)

slaren committed 239 days ago

Verified b47ab7b8

parallel : increase the variability of the prompt lengths (#13927)

ggerganov committed 239 days ago

Verified dd665cc9

Older