ggerganov/llama.cpp

Pull Requests Commits

ggerganov committed 56 days ago

3973163b

ggml: fix ggml_conv_1d_dw bug (ggml/1323)

jason-ni committed 56 days ago

5ade3000

tests : remove unused includes (ggml/0)

ggerganov committed 56 days ago

8b248373

perplexity : provide a helpful hint for has_cpl case in split_equal error. (#15304)

kallewoof committed 56 days ago

Verified 810b9fc8

cuda : fix GGML_CUDA_GRAPHS=OFF (#15300)

CISC committed 57 days ago

Verified 4ebd0c12

finetune: SGD optimizer, more CLI args (#13873)

graehl committed 57 days ago

Verified 5cdb27e0

perplexity: give more information about constraints on failure (#15303)

kallewoof committed 57 days ago

Verified 3ea913f1

HIP: bump requirement to rocm 6.1 (#15296)

IMbackK committed 57 days ago

Verified 29c8fbe4

fix(nix): remove non-functional llama-cpp cachix cache from flake.nix (#15295)

basnijholt committed 57 days ago

Verified 1adc9812

server : enable -td and -tbd parameters (#15172)

CISC committed 57 days ago

Verified b3e16665

ggml : update `ggml_rope_multi` (#12665)

foldl committed 57 days ago

Verified c24f4e26

common : add --override-tensor-draft, --cpu-moe-draft and --n-cpu-moe-draft parameters (#15191)

Copilot committed 57 days ago

Verified d8914fc4

server : filter out harmony thought messages (#15278)

aldehir committed 57 days ago

Verified e885445b

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

alitariq4589 committed 58 days ago

Verified 648ebcdb

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

CISC committed 58 days ago

Verified 07aa869a

ggml : repack block_iq4_nlx8 (#14904)

ggerganov committed 58 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 58 days ago

Verified 6028bf74

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

CISC committed 58 days ago

Verified bc518227

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

Tak-RS committed 58 days ago

Verified e71d48e3

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

IMbackK committed 58 days ago

Verified b0493156

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

Rbiessy committed 58 days ago

Verified f4586ee5

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

rmatif committed 59 days ago

Verified 60a76588

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

am17an committed 59 days ago

Verified efe3a909

CANN: GGML_OP_CPY optimization (#15070)

noemotiovon committed 59 days ago

Verified bbd57b7e

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

yeahdongcn committed 59 days ago

Verified 25ff6f76

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

hipudding committed 59 days ago

Verified be48528b

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

gryffindor-rr committed 59 days ago

Verified cf9e5648

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

ngxson committed 59 days ago

Verified fba5c0d6

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

ngxson committed 59 days ago

Verified 53d0a126

readme : update infra list (#15234)

malzag committed 59 days ago

Verified 27093afe

Newer Older