ggerganov/llama.cpp

Pull Requests Commits

server: add format arg to datetime tool

ngxson committed 110 days ago

d8efdcbd

Update ggml/src/gguf.cpp : Defined virtual keyword for destructor of gguf_writer_base (#25867)

yymin1022 committed 110 days ago

Verified fb92d8f1

common : add support for multiple end sequences in the reasoning budget sampler (#25544)

aldehir committed 110 days ago

Verified 910196f6

tests: synchronize save-load-state generation (#26056)

helanfxz committed 110 days ago

Verified d67c0b41

ui: reduce per-token render cost when streaming (#26053)

Zach Winter committed 111 days ago

Verified 555881eb

ui: remove render effects (#26083)

ServeurpersoCom committed 111 days ago

Verified 96013c51

model: add GLM 5.2 Indexer support (#25407)

pcuenca committed 111 days ago

Verified 88bfee14

ui: fix MCP server display name conflicts in tools lists (#26011)

ServeurpersoCom committed 111 days ago

Verified 95a923a6

server: support "reasoning_effort": "none" in OAI API (#26045)

pnb committed 111 days ago

Verified 27209a59

llama: various bug fixes (#26051)

ngxson committed 111 days ago

Verified 298219f9

HIP: remove rocWMMA FlashAttention (#26046)

JohannesGaessler committed 111 days ago

Verified fa72aecc

opencl: cache compiled cl_program binaries on disk (#26050)

wanghqc committed 111 days ago

Verified ed7adbfe

opencl: do not treat NULL-mask flash attention as causal (#25771)

kumaal committed 111 days ago

Verified 56a83860

skill: create `add-new-model` and `code-review` (#26042)

ngxson committed 111 days ago

Verified 77095ee0

UI: Fix settings precedence, Factory < Admin (--ui-config-file) < Users (Settings panel) (#26002)

ServeurpersoCom committed 111 days ago

Verified 54ce507b

cohere2 moe template parser: enforce JSON schema for text responses if a response schema is provided (#26018)

boondocklabs committed 111 days ago

Verified 8f5ab832

vendor: update subprocess.h (#26061)

ngxson committed 111 days ago

Verified 0cea3622

hexagon: further improved pipeline of the core bits (L2, DMA, MM, FA) (#26049)

max-krasnyansky committed 112 days ago

Verified 0a50d990

hexagon: fix Windows crash when op_poll is enabled (#26029)

adgup committed 112 days ago

Verified c0bc8591

CUDA: fix external compilation of q1_0 MMQ (#25778)

JohannesGaessler committed 112 days ago

Verified 1425386f

args: refactor mlock/mmap/directio into load-mode (#20834)

taronaeo committed 112 days ago

Verified e6dd0e29

contrib: fix leftovers from the AI usage policy update (#26030)

ServeurpersoCom committed 112 days ago

Verified da296d6e

metal : add f16 type support to leaky relu (#25981)

iliailmer committed 113 days ago

Verified c588c4f4

conversion: fix non-MoE NomicBert GGUF conversion error (#25996)

aorko01 committed 113 days ago

Verified d941f6e1

contrib: allow all AI-generated code in general (#26012)

ngxson committed 113 days ago

Verified 4310aa4f

ui: Add a "Default" option for the reasoning selector (#25846)

ServeurpersoCom committed 113 days ago

Verified cf512566

CUDA: Improve NVFP4 W4A4 activation quantization (#25730)

ORippler committed 113 days ago

Verified 1a064ab0

hexagon: activation ops update (#25974)

tboinovski1 committed 113 days ago

Verified 0278d836

mtmd: use RAII for setting and resetting non-causal attention (#25723)

nikwen committed 113 days ago

Verified e0833bf6

feat(ui): add symbolic math support to JS sandbox via nerdamer (#25948)

rankaiyx committed 113 days ago

Verified 61328e6a

Older