ggml-org/llama.cpp

Pull Requests Commits

fix double semicolon

0cc4m committed 48 days ago

df488da9

add coopmat2 support

0cc4m committed 48 days ago

63f85ed2

vulkan: repack q4_0 into aligned arrays

0cc4m committed 49 days ago

a4901b04

metal: add CONV_3D (#19927)

Ra5hidIslam committed 49 days ago

Verified 177c7585

common/autoparser : detect reasoning markers when enable_thinking changes system prompt (#20859)

jhen0409 committed 49 days ago

Verified 7a0b6a63

CANN: add RoPE cache preload before ACL graph capture (#20747)

noemotiovon committed 49 days ago

Verified 07ff0005

fix(openvino): explicit memset in buffer_context allocation (#20857)

thedanhoffman committed 49 days ago

Verified cc18f965

opencl: add flattened Q4_K mv and general Q4_K mm (#20773)

shaofeiqi committed 49 days ago

Verified 84ffd0c1

mtmd: Add dynamic high-resolution image preprocessing for InternVL model (#20847)

bssrdf committed 49 days ago

Verified ec2b787e

mtmd : fix LightOnOCR image preprocessing (#20877)

DorianRudolph committed 49 days ago

Verified d3ac030a

server: allow router to report child instances sleep status (#20849)

ngxson committed 49 days ago

Verified 49bfddec

CUDA: fix BF16 FA compilation (#20865)

JohannesGaessler committed 49 days ago

Verified bd3f1d9d

jinja : refactor token advancement (#20864)

CISC committed 50 days ago

Verified 23c9182c

server: fix Host header (#20843)

kurnevsky committed 50 days ago

Verified 81bc4d3d

support bf16 and quantized type (#20803)

arthw committed 50 days ago

Verified f40a80b4

ggml-cuda: native bf16 flash attention for vec kernel (#20525)

eous committed 50 days ago

Verified db9d8aa4

[CUDA] Increase number of output elements per-thread block if the K-dimension is small (#20635)

gaugarg-nv committed 50 days ago

Verified ccb87fa3

misc : prefer ggml-org models in docs and examples (#20827)

ddh0 committed 50 days ago

Verified 3306dbae

common/grammar: fix grammar parsing issues to prevent stack overflow and hangs (#18604)

aagit committed 50 days ago

Verified 990e4d96

context : use n_embd_out for pooled embedding extraction (#20840)

extfs committed 50 days ago

Verified 212f4521

docs : explicit about banning accounts that violates policy (#19593)

ngxson committed 51 days ago

Verified 568aec82

fix(rpc): prevent division by zero in deserialize_tensor (#20712)

y198nt committed 51 days ago

Verified 2bcdddd5

Convert: Make NVFP4 and MXFP4 HF conversions say NVFP4/MXFP4 instead of BF16 (#20730)

michaelw9999 committed 51 days ago

Verified eac9c6ea

ci : switch from pyright to ty (#20826)

CISC committed 51 days ago

Verified 29b28a98

Add shader count for Intel Arc Pro B60 (#20818)

TheBlueMatt committed 51 days ago

Verified cea560f4

common/parser: fix nasty bug causing subtle corruption of generation prompt (#20825)

pwilkin committed 51 days ago

Verified b1c70e2e

ggml-cpu: add always_inline to tinyBLAS_PPC accumulator saves (#20791)

shalinib-ibm committed 51 days ago

Verified e6ec21e6

ai : limit runtime of the agent (#20816)

ggerganov committed 51 days ago

Verified 4cb7e0bd

common : fix typo in debug log ('extracft' -> 'extract') (#20807)

jpohhhh committed 51 days ago

Verified 149b2493

ai : do not run bash commands in the prompt (#20810)

ggerganov committed 51 days ago

Verified b31b30f3

Older