ggml-org/llama.cpp

Pull Requests Commits
spec: add backend sampling support for eagle3 (#24655)
ruixiang63 committed 58 days ago
Verified a1824902
vulkan: prefer host-visible memory buffers on UMA devices (#22930)
winstonma committed 58 days ago
Verified 32120c10
vulkan: Support gated_delta_net with S_v=16 (#24581)
jeffbolznv committed 58 days ago
Verified d5fb1042
spec: add spec metrics mean acceptance length and acceptance rate per position (#24536)
ruixiang63 committed 58 days ago
Verified 635b65ad
bench : add --offline (#24511)
angt committed 58 days ago
Verified e3a74b29
sycl: support reordered Q4_K/Q5_K/Q6_K MoE MUL_MAT_ID (#24452)
newjordan committed 58 days ago
Verified ac79caa7
[SYCL] Support OP EXPM1, support all UT cases of FLOOR, TRUNC, ROUND (#24363)
arthw committed 58 days ago
Verified fdd10988
sycl : Make GGML_SYCL_F16=ON the default (#23996)
malsbat committed 58 days ago
Verified 4196b477
vulkan: add col2im_1d op (#24425)
ServeurpersoCom committed 58 days ago
Verified ad39ccaa
chat : fix LFM2 tool-call parsing double-escaping (#24667)
tdakhran committed 59 days ago
Verified 7dad2f1a
mtmd: fix miscounting n_tokens (#24656)
ngxson committed 59 days ago
Verified e36a602b
chat: include full unparsed prompt in debug (#24650)
pwilkin committed 59 days ago
Verified 38d54633
docs: Add instructions to install `llama.cpp` from conda-forge (#22219)
jjerphan committed 59 days ago
Verified a1eb756c
chat: harden peg-native tool call parsing (#24329)
ServeurpersoCom committed 59 days ago
Verified 581e8eca
chat: fix an "oldie but goodie" grammar generator bug that surfaced during last changes (#24653)
pwilkin committed 59 days ago
Verified 0ae3f450
mtmd : add post-decode callback (#24645)
ggerganov committed 59 days ago
Verified e3cab403
vulkan: support more CONCAT types (#24579)
jeffbolznv committed 59 days ago
Verified 9dbc6621
wasm : fix fallback symbol collision (#24639)
abetlen committed 59 days ago
Verified 6eab4718
SYCL: use native subgroup size for K-quant DMMV (#21700)
PMZFX committed 59 days ago
Verified e3bb1add
sycl: fix soft_max_f32 max reduction (#24451)
someoneinjd committed 59 days ago
Verified d8a3f523
sycl : fix reorder function; add fp32/fp16 in build script (#24578)
arthw committed 59 days ago
Verified 72be44f1
sycl : enhance set_rows to support q1_0, mxfp4, nvfp4 (#24564)
arthw committed 59 days ago
Verified 8872ab54
[SYCL] add to support pool_1d, move pool_1d/2d code to pool.cpp/hpp (#24584)
arthw committed 59 days ago
Verified 987fbd82
[SYCL]: Remove per-allocation Level Zero runtime checks (#23399)
sanmai committed 59 days ago
Verified c035ff49
metal : add repeat bf16 (#24638)
ggerganov committed 59 days ago
Verified 272088b9
chat: fix whitespace problems once and for all (#24624)
pwilkin committed 59 days ago
Verified a6dff712
UI/svg block rendering (#24080)
ServeurpersoCom committed 59 days ago
Verified 2a6c391a
CUDA: only support F32/F16 for GGML_OP_REPEAT (#24533)
leonardHONG committed 59 days ago
Verified 3686e9d6
ggml-webgpu: improve i-quants mul_mat performance and speed up prefill (#24530)
yomaytk committed 59 days ago
Verified 6e9007ae
convert : fix lora base model arch retrieval (#24621)
CISC committed 60 days ago
Verified dd4623a7
Terms Of Use Privacy Policy Refund Policy Impressum