ggml-org/llama.cpp

Pull Requests Commits

Update tools/llama-bench/llama-bench.cpp

max-krasnyansky committed 233 days ago

Verified ac35e50c

threading: disable SetThreadInfo() calls for older Windows versions

max-krasnyansky committed 234 days ago

9087dd26

threading: support for GGML_SCHED_PRIO_LOW, update thread info on Windows to avoid throttling

max-krasnyansky committed 234 days ago

199a8384

CUDA: fix typo in FlashAttention code (#13926)

JohannesGaessler committed 234 days ago

Verified e562eece

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

slaren committed 234 days ago

Verified b47ab7b8

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

ggerganov committed 234 days ago

Verified dd665cc9

cuda : prevent using split buffers with 3d/4d matrices (#13919)

slaren committed 234 days ago

Verified df0c0c7d

SYCL: Add mrope kernel (#13755)

qnixsynapse committed 234 days ago

Verified b49a8ff9

sync : vendor (#13901)

ggerganov committed 234 days ago

Verified 53f92507

convert : fix rwkv bos/eos token (#13844)

CISC committed 234 days ago

Verified db38704f

convert : allow partial update to the chkhsh pre-tokenizer list (#13847)

ngxson committed 234 days ago

Verified 07e4351c

llama : add support for DistilBert (#13907)

huydt84 committed 234 days ago

Verified 291f2b69

llama : use llm_build_granite for minicpm (#13911)

zkh2016 committed 234 days ago

Verified 2c90da4c

cmake: Guard GGML_CPU_ALL_VARIANTS by architecture (#13890)

ckastner committed 235 days ago

Verified ec9e0301

llama : add support for jina-reranker-v2 (#13900)

CISC committed 235 days ago

Verified e83ba3e4

gguf-py : add support for sub_type (in arrays) in GGUFWriter add_key_value method (#13561)

CISC committed 235 days ago

Verified 2b131621

arm64: optimize q4_k_q8_k kernel with i8mm (#13886)

cyb70289 committed 235 days ago

Verified 54a2c7a8

cmake: Factor out CPU architecture detection (#13883)

ckastner committed 235 days ago

Verified 21fcc21a

ggml: aarch64: Implement SVE F32 kernels for Mamba Sequential Scan Algorithm (#13882)

vineelabhinav committed 235 days ago

Verified dd8ba934

tests : remove json.hpp from a test (#13880)

ggerganov committed 235 days ago

Verified 66c92061

convert : workaround for AutoConfig dummy labels (#13881)

CISC committed 235 days ago

Verified 5ca82fc1

llama : add RobertaForSequenceClassification reranker support (#13875)

CISC committed 235 days ago

Verified 6385b843

ggml: aarch64: Implement SVE F32 kernels for vector functions (#13843)

vineelabhinav committed 235 days ago

Verified 1b8fb815

gguf-py : fix SafetensorRemote return on undefined size (< 0) (#13841)

Beinsezii committed 236 days ago

Verified 53ae3064

llama : fix KV shift for qwen2vl (#13870)

ngxson committed 236 days ago

Verified 763d06ed

mtmd : move helpers to dedicated library (⚠️ breaking change) (#13866)

ngxson committed 236 days ago

Verified 10961339

ci: disable LLAMA_CURL for Linux cross-builds (#13871)

mtmcp committed 236 days ago

Verified d98f2a35

llama : add support for BertForSequenceClassification reranker (#13858)

huydt84 committed 236 days ago

Verified e0e3aa23

convert: small addition to support LlamaModel (#13838)

huydt84 committed 236 days ago

Verified aa6dff05

server: fix remove 'image_url'/'input_audio' json-object effectlly for 'llama_params' in multimodal-model-mode (#13853)

flyinskyin2013 committed 236 days ago

Verified c962ae33

Older