ngxson/llama.cpp

Pull Requests Commits
llama-bench : clarify benchmarked parts of the computation (#16823)
ggerganov committed 217 days ago
Verified a8ca18b4
initialise buffer.device in ggml_hexagon_session (#16816)
l3utterfly committed 217 days ago
Verified 8284efc3
embedding: add raw option for --embd-output-format (#16541)
SamMalayek committed 218 days ago
Verified 1c1409e1
llama: consistent ctx <-> buf order for KV cache (#16746)
JohannesGaessler committed 218 days ago
Verified 7a0e900e
grammar : support array references in json schema (#16792)
aldehir committed 218 days ago
Verified 280d97be
CANN: Improve device ID handling and aclnnArange checks (#16752)
noemotiovon committed 218 days ago
Verified 3479efd1
CUDA: add unused vars to mmvf and mmvq (#16807)
am17an committed 218 days ago
Verified 463bbf20
sycl: add SSM_CONV operation support (#16800)
tamarPal committed 218 days ago
Verified ad8d36be
chat: Add LFM2 tool handling (#16763)
ykhrustalev committed 218 days ago
Verified c053e18a
mtmd : fix idefics3 preprocessing (#16806)
ngxson committed 218 days ago
Verified e1ab0848
llama : disable pipeline parallelism if compute buffer allocation fails (#16748)
slaren committed 218 days ago
Verified 5a4ff43e
ggml : fix interpolate with align-corners and ne=1 (#16700)
Acly committed 218 days ago
Verified 10640e31
HIP: fix AMDGPU_TARGETS, update documentation (#16803)
JohannesGaessler committed 218 days ago
Verified 80d28f10
model : add LightOnOCR-1B model (#16764)
ngxson committed 218 days ago
Verified c55d53ac
llama: fix leaked buffers for mmap + split files (#16765)
JohannesGaessler committed 219 days ago
Verified 945501f5
test-backend-ops: print failed tests at the end (#16785)
am17an committed 219 days ago
Verified 75cbdd3f
sycl: add ROLL operation support (#16665)
tamarPal committed 219 days ago
Verified 2b9bd9bf
sycl: add REPEAT_BACK operation support (#16734)
shani-f committed 219 days ago
Verified 59fc1ec8
CUDA: support for weight clamp in top-k norm (#16702)
am17an committed 219 days ago
Verified 75d33b93
ggml-alloc : make gallocr prefer chunks that allow memory reuse (#16788)
Acly committed 219 days ago
Verified 3470a5c8
cuda : use fast copy when src and dst are of different type and contiguous (#16789)
CISC committed 219 days ago
Verified bd562fe4
ggml: fix cuda kernel launch configuration for k_compute_batched_ptrs to support large batch (#16744)
leejet committed 219 days ago
Verified bbac6a26
convert : enable expert group selection for all models with it (#16691)
CISC committed 219 days ago
Verified 73a48c97
graph : add clamping to ffn_moe_weights_sum to avoid div-by-zero (#16655)
CISC committed 219 days ago
Verified f696428c
model : set res->t_embd in SmallThinker models (#16782)
CISC committed 219 days ago
Verified 7cce4f81
docs : add Jamba to Text-only models list (#16778)
amirai21 committed 220 days ago
Verified 8d886282
CUDA: General GEMV fusion (#16715)
am17an committed 220 days ago
Verified f77c13b9
vulkan: deduplicate Microsoft Direct3D12 devices (#16689)
giladgd committed 220 days ago
Verified 3cfa9c3f
convert : handle mmproj filename/path properly (#16760)
Galunid committed 220 days ago
Verified 5d195f17
model : set res->t_embd in PLaMo2 models (#16766)
mitmul committed 221 days ago
Verified 226f295f
Terms Of Use Privacy Policy Refund Policy Impressum