LLM inference in C/C++
C++ ggml
Commits on master
sycl : Enable to support fp16 by OPs: SQR, SQRT, LOG, SIN, COS, CLAMP (#24692)
58728bd Neo Zhang committed 5 months ago
SYCL: fix use-after-free bug with async memcpy in MoE prefill (#24676)
ebbc1e5 Alexey Kopytko committed 5 months ago
vendor : update BoringSSL to 0.20260616.0 (#24693)
74ade52 Alessandro de Oliveira Faria (A.K.A.CABELO) committed 5 months ago
Fix and restrict NVFP4 edge-cases in llama-graph (#24331)
02810c7 Oliver Simons committed 5 months ago
vulkan: prefer host-visible memory buffers on UMA devices (#22930)
32120c1 Winston Ma committed 5 months ago
spec: add spec metrics mean acceptance length and acceptance rate per position (#24536)
635b65a Ruixiang Wang committed 5 months ago
sycl: support reordered Q4_K/Q5_K/Q6_K MoE MUL_MAT_ID (#24452)
ac79caa Frosty40 committed 5 months ago
[SYCL] Support OP EXPM1, support all UT cases of FLOOR, TRUNC, ROUND (#24363)
fdd1098 Neo Zhang committed 5 months ago
chat : fix LFM2 tool-call parsing double-escaping (#24667)
7dad2f1 Tarek Dakhran committed 5 months ago
chat: include full unparsed prompt in debug (#24650)
38d5463 Piotr Wilkin (ilintar) committed 5 months ago
docs: Add instructions to install `llama.cpp` from conda-forge (#22219)
a1eb756 Julien Jerphanion committed 5 months ago
chat: fix an "oldie but goodie" grammar generator bug that surfaced during last changes (#24653)
0ae3f45 Piotr Wilkin (ilintar) committed 5 months ago
sycl : fix reorder function; add fp32/fp16 in build script (#24578)
72be44f Neo Zhang committed 5 months ago
sycl : enhance set_rows to support q1_0, mxfp4, nvfp4 (#24564)
8872ab5 Neo Zhang committed 5 months ago
[SYCL] add to support pool_1d, move pool_1d/2d code to pool.cpp/hpp (#24584)
987fbd8 Neo Zhang committed 5 months ago
[SYCL]: Remove per-allocation Level Zero runtime checks (#23399)
c035ff4 Alexey Kopytko committed 5 months ago