ggml-org/llama.cpp

Pull Requests Commits

CUDA: dequant quantized kv on-the-fly for vec kernels

am17an committed 36 days ago

b9e92c43

vulkan: extend sparse FA support to coopmat2 (#30003)

jeffbolznv committed 37 days ago

Verified dac30873

vendor : update cpp-httplib to 0.60.1 (#30134)

angt committed 37 days ago

Verified 03aa006a

cuda : support arbitrary striding for unary ops on f16, f32, and bf16 (#29781)

saady789 committed 37 days ago

Verified 08246a28

sycl: FWHT optimizations (#29605)

bri-prism committed 37 days ago

Verified 46baf1f1

hex-mmadd: do not assume aligned read/write when bias-add is fused (#30133)

max-krasnyansky committed 37 days ago

Verified 097f5b53

cuda : Use byte strides for roll to allow non-contiguous ROLL operations (#29547)

bertaye committed 37 days ago

Verified d8880160

CUDA: fix norm family kernels when ne[2]/ne[3] exceed grid dim limits (#28175)

edenfunf committed 37 days ago

Verified fda18666

sycl: fuse the delta-net alpha gate (add + unary + mul) (#29687)

Titaniumtown committed 37 days ago

Verified ff30363a

sycl: stage bulk uploads (model loading) through a pinned ring buffer (#29608)

cwriter committed 37 days ago

Verified 000bee54

model : support classifier_activation for rerankers (#29692)

boshjerns committed 37 days ago

Verified 37ac6345

ggml-cuda: assign four GDN state columns per warp (#30087)

SongXiaoXi committed 37 days ago

Verified 24e41838

sycl: add grouped MoE XMX GEMM (#29245)

cwriter committed 37 days ago

Verified 847f447c

convert : support Qwen3.5 embedding models (#27920)

SamMalayek committed 37 days ago

Verified 75118a3a

sycl: remove duplicate block-size defines from op headers (#29507)

Titaniumtown committed 37 days ago

Verified 9b4ed0ca

hexagon: enable alloc_buffer_n (#30126)

max-krasnyansky committed 37 days ago

Verified 9c2e0e49

hexagon: Q6_K weight dequant speedup (#30121)

kurquhar committed 37 days ago

Verified 06cad0b9

model : add LiquidAI/d1-omni-600M decision model (#30114)

tdakhran committed 37 days ago

Verified a657f7e9

hexagon: support tiled Q4_K and Q6_K GET_ROWS (#30115)

kurquhar committed 37 days ago

Verified aa5e0092

hexagon: improved GELU accuracy (#30104)

kurquhar committed 37 days ago

Verified 70815103

chat: fix jinja parser for TranslateGemma (#30096)

nbud committed 37 days ago

Verified bd4eeaa0

chat : name tool and argument parser rules by index (#30088)

Frost-54 committed 37 days ago

Verified 5de73343

model : add LiquidAI/d1-3B decision model (#30110)

tdakhran committed 37 days ago

Verified 88dcc460

cuda: FWHT kernels for block widths above 512 (#29100)

bri-prism committed 37 days ago

Verified b86d2f07

mtmd: add cohere2 vision support (#30062)

Terrencezzj committed 37 days ago

Verified 50a6c5cf

llama : add a GPU cache for MoE experts kept in host memory (#29887)

am17an committed 37 days ago

Verified d6cf9acb

cuda: update uncoalesced memory reads in pool2d (#29425)

shenron0101 committed 37 days ago

Verified 42c787e8

server : accumulate generated text and tokens as parse input (#29876)

aldehir committed 37 days ago

Verified 18b5f8b1

llama: share the nextn tensor flags between models (#30097)

ServeurpersoCom committed 37 days ago

Verified 448147d4

metal : few-row MMA mat-mul for the remaining src0 types (#30065)

pratiknarola-t committed 37 days ago

Verified 98819068

Older