ggml-org/llama.cpp

Pull Requests Commits

llama-eval : add AIME 2026 dataset support

ggerganov committed 10 hours ago

Verified 28450403

webui: Move static build output from repo code to HF Bucket (#22937)

allozaur committed 13 hours ago

Verified 253ba110

logs : reduce (#23021)

ggerganov committed 14 hours ago

Verified 67b2b7f2

ggml-cpu: Add IME2 Instruction Support for the SpacemiT Backend (#22863)

alex-spacemit committed 15 hours ago

Verified 81b0d882

docker : revert stable version of intel compute-runtime (#22968)

arthw committed 15 hours ago

Verified 0f45f1a3

unicode,test: add Qwen3.5 non-backtracking tokenizer handler and regr… (#22110)

Kabir08 committed 15 hours ago

Verified 42532aff

vulkan: fix matmul integer pipeline selection (#23005)

0cc4m committed 16 hours ago

Verified dbe7901c

fix: Autoscroll detection (#23026)

allozaur committed 18 hours ago

Verified 320a6a44

SYCL: fix multi-GPU system RAM exhaustion by using Level Zero allocations (#21597)

PMZFX committed 19 hours ago

Verified 9ed6e19b

ggml-webgpu: only use subgroup-matrix path when head dims are divisible by sg_mat_k / sg_mat_n (#23020)

ArberSephirotheca committed 1 day ago

Verified 4c1c3ac0

Fix for issue #22974. Cast intermediate results to float before adding and casting the result to the destination type. Avoids half+half operator ambiguity. (#22994)

scutler-nv committed 1 day ago

Verified 7f3f843c

opencl: add q5_0 and q5_1 MoE for Adreno (#22985)

shaofeiqi committed 1 day ago

Verified ec562eb6

server, webui: accept continue_final_message flag for vLLM API compat (#23012)

ServeurpersoCom committed 1 day ago

Verified 95d469a9

opencl: fix crash when warming up MoE on Adreno (#22876)

lhez committed 1 day ago

Verified 1e4579fb

flush the gpu profile timestamp before the queryset is overflowed (#22995)

yomaytk committed 1 day ago

Verified 527045bf

webui: Deduplicate model aliases in data + handle single/multiple aliases in UI (#22979)

allozaur committed 1 day ago

Verified 2dfeca31

webui: preserve system message on edit cancel (#22911)

ServeurpersoCom committed 1 day ago

Verified 46be24d1

docs : Update OPENVINO.md (#22959)

ravi9 committed 1 day ago

Verified 7e166460

hexagon: add unary tanh op (#22999)

max-krasnyansky committed 1 day ago

Verified ad96bb8c

download: do not exit() on error (#23008)

ngxson committed 1 day ago

Verified e75cd5ef

server, webui: support continue generation on reasoning models (#22727)

ServeurpersoCom committed 1 day ago

Verified 5d44db60

ci: validate model naming convention (#22680)

ngxson committed 1 day ago

Verified 3796c94b

spec : update CLI arguments for better consistency (#22964)

ggerganov committed 1 day ago

Verified 634275fb

llama-eval : enable type check (#22988)

CISC committed 1 day ago

Verified bcfe63fc

ggml-zendnn : adaptive fallback to CPU backend for small batch sizes (#22681)

z-sachin committed 1 day ago

Verified 61af07c2

hexagon: eliminate scalar VTCM loads via HVX splat helpers (#22993)

trivikram-reddy1 committed 2 days ago

Verified 856c3ada

opencl: add opt-in Adreno xmem F16xF32 GEMM for prefill (#22755)

happyyzy committed 2 days ago

Verified a9883db8

convert : fix Pixtral 12B --mistral-format conversion (3 bugs) (#22981)

fredzillman committed 2 days ago

Verified cce09f0b

webui: Fix Chat Screen Form box disappearing + autoscroll issues on WebKit (#22977)

allozaur committed 2 days ago

Verified dded58b4

mtmd, server, common: expose modalities to /v1/models (#22952)

ngxson committed 2 days ago

Verified 7bfe120c

Older