ngxson/llama.cpp

Pull Requests Commits
opencl: add `mul_mv_id_q4_0_f32_8x_flat` (#14003)
lhez committed 267 days ago
Verified 4c763c8d
kv-cache : avoid modifying recurrent cells when setting inputs (#13834)
compilade committed 267 days ago
Verified dad5c443
convert : fix duplicate key DeepSeek-R1 conversion error (#14103)
CISC committed 267 days ago
Verified 55f6b9fa
llama : support GEGLU for jina-bert-v2 (#14090)
CISC committed 267 days ago
Verified 3678b838
vulkan: force device 0 in CI (#14106)
jeffbolznv committed 267 days ago
Verified 652b70e6
Fixed spec timings to: accepted/tested instead of accepted/drafted (#14104)
jukofyork committed 267 days ago
Verified 3a12db23
sync : ggml
ggerganov committed 267 days ago
ae92c185
ggml : fix weak alias win32 (whisper/0)
ggerganov committed 267 days ago
b7ce1ad1
Vulkan: Don't default to CPU device (like llvmpipe), even if no other device is available, to allow fallback to CPU backend (#14099)
0cc4m committed 268 days ago
Verified 97340b4c
rpc : nicer error messages for RPC server crash (#14076)
isaac-mcfadyen committed 268 days ago
Verified 2bb04670
sync : ggml
ggerganov committed 268 days ago
b8e2194e
Add in-build ggml::ggml ALIAS library (ggml/1260)
dg0yt committed 268 days ago
1a3b5e80
metal : use less stack memory in FA kernel (#14088)
ggerganov committed 268 days ago
Verified 1f63e75f
kv-cache : fix shift and defrag logic (#14081)
ggerganov committed 268 days ago
Verified 40cbf571
llama : allow building all tests on windows when not using shared libs (#13980)
slaren committed 268 days ago
Verified 7f4fbe51
ggml-cpu : split arch-specific implementations (#13892)
xctan committed 268 days ago
Verified f470bc36
cuda : fix device sync on buffer clear (#14033)
slaren committed 268 days ago
Verified 8f47e25f
graph : fix geglu (#14077)
ggerganov committed 268 days ago
Verified 201b31dc
CANN: Simplify the environment variable setting(#13104)
bachelor-dou committed 269 days ago
Verified e21d2d4a
webui: fix sidebar being covered by main content (#14082)
yeahdongcn committed 269 days ago
Verified dc0623fd
server : fix LRU check (#14079)
ggerganov committed 269 days ago
Verified 87d34b38
sycl: Add reorder to Q6_K mmvq implementation (#13885)
s-Nick committed 269 days ago
Verified b460d16a
add geglu activation function (#14074)
huydt84 committed 269 days ago
Verified 91a8ee6a
CANN: Enable labeler for Ascend NPU (#13914)
shink committed 269 days ago
Verified 056eb745
cuda : fix buffer type check with integrated GPUs (#14069)
slaren committed 269 days ago
Verified 247e5c6e
ci: add LoongArch cross-compile build (#13944)
wojiushixiaobai committed 271 days ago
Verified 5787b5da
SYCL: Implement few same quantized type copy kernels (#13739)
qnixsynapse committed 271 days ago
Verified 228f34c9
llama : fix llama_model_chat_template with template name (LLM_KV with suffix) (#14050)
CISC committed 271 days ago
Verified 0974ad7a
llama : deprecate llama_kv_self_ API (#14030)
ggerganov committed 272 days ago
Verified 745aa531
context : fix SWA-related warning for multiple sequences (#14045)
ggerganov committed 272 days ago
Verified 487a5e04
Terms Of Use Privacy Policy Refund Policy Impressum