ngxson/llama.cpp

Pull Requests Commits
Reduce level of content parser warning message to avoid log spam on non-debug verbosity (#20347)
pwilkin committed 206 days ago
Verified 6c770d16
examples : fix empty items in json_schema_to_grammar.py [no ci] (#19968)
RayXu14 committed 206 days ago
Verified 8d880ac0
docs: update CPU backend ops to mark POOL_1D as supported (#20304)
a3894281 committed 206 days ago
Verified 0f1e9d14
models : fix assert in mamba2 (cont) (#20335)
ggerganov committed 206 days ago
Verified 1274fbee
server : make 2 checkpoints near the end of the prompt (#20288)
ggerganov committed 206 days ago
Verified a7b3dee7
common : fix incorrect uses of stoul (#20313)
CISC committed 207 days ago
Verified ec947d2b
kleidiai : support for concurrent sme and neon kernel execution (#20070)
chaxu01 committed 207 days ago
Verified 0cd4f472
ggml-cpu: add RVV repack GEMM and GEMV for quantization types (#19121)
taimur-10x committed 207 days ago
Verified af237f30
metal: handle command buffer failures gracefully in synchronize (#20306)
JulianPscheid committed 207 days ago
Verified 1a5631be
llama-quant : fail early on missing imatrix, refactor type selection, code cleanup (#19770)
ddh0 committed 207 days ago
Verified 1dab5f5a
common: consolidate PEG string parsers (#20263)
aldehir committed 207 days ago
Verified c96f608d
model: fix step3.5 n_rot (#20318)
ngxson committed 207 days ago
Verified 0842b9b4
llama: dynamic head_dim and n_rot for SWA (#20301)
ngxson committed 207 days ago
Verified 59db9a35
server: Parse port numbers from MCP server URLs in CORS proxy (#20208)
eapache committed 207 days ago
Verified 23fbfcb1
metal : extend mul_mv_ext to BF16, Q2_K, Q3_K (#20250)
arkavo-com committed 207 days ago
Verified e22cd0aa
server : fix checkpoints n_tokens calculation (#20287)
ggerganov committed 207 days ago
Verified 96cfc499
metal : add upscale (#20284)
ggerganov committed 207 days ago
Verified ed0007aa
server : warn swa-full is not supported for non-SWA models (#20291)
ggerganov committed 207 days ago
Verified 344ee2a3
server : fix off-by-1 in server_tokens::size_up_to_pos() (#20279)
ggerganov committed 207 days ago
Verified d6e15564
common: map developer role to system (#20215)
pwilkin committed 207 days ago
Verified f76565db
models : fix assert in mamba2 graph (#20270)
ggerganov committed 208 days ago
Verified 43e1cbd6
server : add kill switch when server is stuck (#20277)
ggerganov committed 208 days ago
Verified 107d5999
ggml-cuda: disable gdn for musa (#20278)
am17an committed 208 days ago
Verified e8bbc736
llama-quant : left-align tensor names in output (#20117)
ddh0 committed 208 days ago
Verified b5181951
contributing: limit open PRs for new contributors to 1 (#20036)
am17an committed 208 days ago
Verified e2763a67
ggml-vulkan: add SGN operator, auto-generate Vulkan.csv and ops.md (#20219)
bertaye committed 208 days ago
Verified 0beb8db3
vulkan: skip zero size tensors in backend copies (#20233)
0cc4m committed 208 days ago
Verified b2f460bd
cuda : display total and free VRAM capacity during device initialization (#20185)
tehsiuhuang committed 208 days ago
Verified 5f4cdac3
llama-bench: introduce `-hf` and `-hff` flags & use `--mmap 1` by default (#20211)
taronaeo committed 208 days ago
Verified ae87863d
PEG parser for LFM2 (#20251)
pwilkin committed 208 days ago
Verified 97c64fbd
Terms Of Use Privacy Policy Refund Policy Impressum