ngxson/llama.cpp

Pull Requests Commits

wip lfm2 vision model

ngxson committed 176 days ago

955b18e9

server : filter out harmony thought messages (#15278)

aldehir committed 176 days ago

Verified e885445b

ci : Added CI with RISC-V RVV1.0 Hardware (#14439)

alitariq4589 committed 176 days ago

Verified 648ebcdb

ci : add more python requirements to copilot-setup-steps (#15289)

CISC committed 176 days ago

Verified 07aa869a

ggml : repack block_iq4_nlx8 (#14904)

ggerganov committed 176 days ago

Verified 00f35d50

CUDA: Optimize `reduce_rows_f32` kernel, leading up to 25x perf improvement on kernel-level and 10% perf increase for Gemma3n (#15132)

ORippler committed 176 days ago

Verified 6028bf74

ci : add copilot-setup-steps.yml (#15214)

CISC committed 176 days ago

Verified bc518227

ggml-rpc: chunk send()/recv() to avoid EINVAL for very large tensors over RPC (macOS & others) (#15188)

Tak-RS committed 177 days ago

Verified e71d48e3

HIP: disable sync warp shuffel operators from clr amd_warp_sync_functions.h (#15273)

IMbackK committed 177 days ago

Verified b0493156

sycl: Fix and disable more configurations of mul_mat (#15151)

Rbiessy committed 177 days ago

Verified f4586ee5

opencl: allow mixed f16/f32 `add` (#15140)

rmatif committed 177 days ago

Verified 60a76588

CUDA cmake: add `-lineinfo` for easier debug (#15260)

am17an committed 177 days ago

Verified efe3a909

CANN: GGML_OP_CPY optimization (#15070)

noemotiovon committed 177 days ago

Verified bbd57b7e

musa: fix failures in test-backend-ops for mul_mat_id op (#15236)

yeahdongcn committed 178 days ago

Verified 25ff6f76

CANN: Add broadcast for softmax and FA (#15208)

hipudding committed 178 days ago

Verified be48528b

mtmd : Fix MinicpmV model converter and clip to avoid using hardcode. (#14750)

gryffindor-rr committed 178 days ago

Verified cf9e5648

chat : hotfix gpt-oss jinja raising an exception (#15243)

ngxson committed 178 days ago

Verified fba5c0d6

server : allow specifying reasoning_format in HTTP request (#15238)

ngxson committed 178 days ago

Verified 53d0a126

readme : update infra list (#15234)

malzag committed 178 days ago

Verified 27093afe

kv-cache : fix seq_rm with seq_id == -1 (#15226)

ggerganov committed 178 days ago

Verified 228f724d

kv-cache : log (debug) all streams in find_slot (#15176)

danbev committed 178 days ago

Verified cd3069df

convert : fix merge conflicts (#15229)

CISC committed 178 days ago

Verified 50e81bdf

perplexity : update comments/error msg to use decode [no ci] (#15227)

danbev committed 178 days ago

Verified 1ebbaddf

convert : improve Mistral models integration (#14737)

juliendenize committed 178 days ago

Verified a3a78742

kleidiai: fix unsigned overflow bug (#15150)

chaxu01 committed 178 days ago

Verified 002cb1bb

cuda: refactored ssm_scan and use CUB (#13291)

Your-Cheese committed 180 days ago

Verified 79c1160b

CUDA: add attention sinks for tile and wmma (#15178)

am17an committed 180 days ago

Verified 34c9d765

gguf-py : add Numpy MXFP4 de/quantization support (#15111)

compilade committed 181 days ago

Verified e54d41be

server-bench: external OAI servers, sqlite (#15179)

JohannesGaessler committed 181 days ago

Verified 4850b52a

ggml : fix field name when new ggml_backend (#14944)

aisk committed 181 days ago

Verified cd6983d5

Older