ggml-org/llama.cpp

Pull Requests Commits

fit : wrap llama_device_memory_data

ggerganov committed 26 days ago

Verified 35180618

fit : avoid including llama-ext.h in fit.h (#24506)

ggerganov committed 26 days ago

Verified 02182fc5

ggerganov committed 26 days ago

f532be8f

ggml : bump version to 0.15.1 (ggml/1541)

ggerganov committed 26 days ago

e08c226a

vendor : update cpp-httplib to 0.47.0 (#24395)

angt committed 27 days ago

Verified 70b54e14

UI/jpeg exif orientation (#24196)

ServeurpersoCom committed 27 days ago

Verified 6471e3c0

spec: add EAGLE3 speculative decoding support (#18039)

ruixiang63 committed 27 days ago

Verified 88a39274

ggml: support concat for scalar types at cuda backend (#24011)

zihaomu committed 27 days ago

Verified 85f99dca

[SYCL] Fix CI build & release for SYCL backend (#24387)

arthw committed 27 days ago

Verified 099ea76f

opencl: add q5_0/q5_1 gemm and gemv kernels for Adreno (#24319)

shaofeiqi committed 27 days ago

Verified ba1df050

docker : support specifying the GCC version for CUDA (#24447)

wencan committed 27 days ago

Verified 1593d568

vulkan: ifdef eMesaHoneykrisp (build fix) (#24479)

jeffbolznv committed 27 days ago

Verified 4c659550

ggerganov committed 27 days ago

263cc04a

ggml : bump version to 0.15.0 (ggml/1539)

ggerganov committed 27 days ago

17e59d62

vulkan: add fast path for contiguous buffer transfers (#23973)

winstonma committed 27 days ago

Verified fdc3db9b

vulkan: use medium matmul tile on Asahi Linux (#24306)

xingjianll committed 27 days ago

Verified 1af154a7

server: skip unused log lines on router mode (#24463)

ngxson committed 28 days ago

Verified 18ef86ec

vocab : adopt leading TemplateProcessing special token as BOS (#24428)

o7si committed 28 days ago

Verified 1bfbdb13

vocab : refactor normalizer flags into options struct, add strip_accents (#24371)

o7si committed 28 days ago

Verified 68f30663

server : skip checkpoints beyond pos_next (#24411)

aldehir committed 28 days ago

Verified db94854f

vendor : update LibreSSL to 4.3.2 (#24397)

angt committed 28 days ago

Verified ac4cddeb

Remove padding and multiple D2D copies for MTP (#24086)

gaugarg-nv committed 28 days ago

Verified e95dae18

chat: fix LFM2/LFM2.5 ignoring json_schema (#24377)

tdakhran committed 28 days ago

Verified d2462f8f

CUDA: Fix ssm_scan_f32 data-races (#24360)

ORippler committed 28 days ago

Verified fb83cc9a

ci : bump komac version (#24396)

CISC committed 29 days ago

Verified 039e20a2

speculative : fix "ngram-map-k4v" name in logging (#24253)

ddh0 committed 29 days ago

Verified d2e22ed9

webui: implement pinned conversations support (#21387)

remeh committed 29 days ago

Verified 76da2450

graph: Fix granite speech model inference by applying embedding scale when deepstack is not used (#24357)

arnu515 committed 29 days ago

Verified d73cd076

ci : fix windows release (#24369)

CISC committed 29 days ago

Verified e25a32e9

ui: add opt-in run_javascript frontend tool (#24244)

ServeurpersoCom committed 29 days ago

Verified 48360950

Older