ggml-org/llama.cpp

Pull Requests Commits

server : fix appearance of the chats list context menu for Safari (#14322)

rntk committed 170 days ago

Verified 83790b0e

SYCL: disable faulty fp16 exp kernel (#14395)

qnixsynapse committed 170 days ago

Verified f47c1d71

ggml : fix unmerged GGML_FPxx_TO_FPxx refactoring (#14443)

CISC committed 170 days ago

Verified a5d1fb62

ggml : implement REGLU/GEGLU/SWIGLU ops (#14158)

CISC committed 170 days ago

Verified a0535ffa

vulkan: Add fusion support for RMS_NORM+MUL (#14366)

jeffbolznv committed 170 days ago

Verified bd9c981d

CUDA: add bf16 and f32 support to cublas_mul_mat_batched (#14361)

am17an committed 171 days ago

Verified 27208bf6

vulkan: handle noncontig in the final case of ggml_vk_get_cpy_pipeline (#14378)

jeffbolznv committed 171 days ago

Verified 63a7bb3c

vulkan: lock accesses of pinned_memory vector (#14333)

jeffbolznv committed 171 days ago

Verified 00d5282c

model : add support for ERNIE 4.5 0.3B model (#14408)

ownia committed 171 days ago

Verified 566c16fc

fix async_mode bug (#14432)

bachelor-dou committed 171 days ago

Verified b25e9277

ci : fix windows build and release (#14431)

CISC committed 171 days ago

Verified 6609507a

vulkan: Fix GGML_VULKAN_SHADER_DEBUG_INFO (#14427)

jeffbolznv committed 171 days ago

Verified ceb1bf5a

graph : make llm_graph_context destructor virtual (#14410)

ggerganov committed 172 days ago

Verified 72babea5

recurrent : call balloc split_reset() in init_batch() (#14414)

ggerganov committed 172 days ago

Verified 43678060

ggml : add ggml_set_rows (#14274)

rgerganov committed 172 days ago

Verified 8d94219a

convert : fix broken sentencepiece vocab (#14416)

CISC committed 172 days ago

Verified f667f1e6

model : gemma3n text-only (#14400)

ngxson committed 173 days ago

Verified 8846aace

cmake: regen vulkan shaders when shaders-gen sources change (#14398)

bandoti committed 173 days ago

Verified a01047b0

llama : return mistral-v7-tekken as default template only (#14390)

CISC committed 173 days ago

Verified b2534622

metal : add special-case mat-vec mul for ne00 == 4 (#14385)

ggerganov committed 173 days ago

Verified e8215dbb

metal : batch rows copy in a single threadgroup (#14384)

ggerganov committed 173 days ago

Verified 5783ae43

docs: update s390x documentation + add faq (#14389)

taronaeo committed 173 days ago

Verified bf5bcd0b

musa: enable fp16 mma (all) and cublas on qy2 (#13842)

yeahdongcn committed 173 days ago

Verified 716301d1

ggml-cpu: enable IBM NNPA Vector Intrinsics (#14317)

taronaeo committed 174 days ago

Verified 60ef23d6

ggml : do not output unprintable characters on GGUF load failure (#14381)

CISC committed 174 days ago

Verified b193d530

sycl: GGML_SYCL_DISABLE_OPT on by default for all Intel Devices (#13973)

ShanoToni committed 174 days ago

Verified 2bf9d539

opencl: ref count `ggml_backend_opencl_context` and refactor profiling (#14254)

lhez committed 175 days ago

Verified 73e53dc8

batch : fix check for empty sequences in memory (#14364)

ggerganov committed 175 days ago

Verified 62af4642

cmake : use LLAMA_BUILD_NUMBER when defining LLAMA_INSTALL_VERSION (#14362)

mbaudier committed 175 days ago

Verified c148cf19

server : move no API key doc to /health (#14352)

pnb committed 175 days ago

Verified 1b809cee

Newer Older