ggml-org/llama.cpp

Pull Requests Commits
ggml-cpu : optimize RVV q2_k and q3_k kernels (#16887)
xctan committed 47 days ago
Verified 7f09a680
CUDA: fix crash on uneven context without FA (#16988)
JohannesGaessler committed 47 days ago
Verified aa374175
metal : initial Metal4 tensor API support (#16634)
ggerganov committed 47 days ago
Verified 5b180c3d
server : disable checkpoints with mtmd (#17045)
ggerganov committed 47 days ago
Verified b7f9010d
clip: implement minicpm-v sinusoidal embd using GGML (#17036)
ngxson committed 47 days ago
Verified 4882f0ff
sycl: add CONCAT operator support (#16047)
ye-NX committed 47 days ago
Verified 9d7c518d
docs: explain CUDA 11 compilation [no ci] (#16824)
JohannesGaessler committed 47 days ago
Verified 22c8c3c6
ggml-hexagon: graceful fallback for older socs where rpcmem_alloc2 and FASTRPC_GET_URI is unsupported (#16987)
l3utterfly committed 47 days ago
Verified 6db3d1ff
improve CUDA cpy memory bandwidth when copying transposed tensor (#16841)
bssrdf committed 47 days ago
Verified 230d1169
vulkan: Fix GGML_VULKAN_CHECK_RESULTS to better handle fusion (#16919)
jeffbolznv committed 48 days ago
Verified a44d7712
examples(gguf): GGUF example outputs (#17025)
gabe-l-hart committed 48 days ago
Verified 5886f4f5
mtmd: allow QwenVL to process larger image by default (#17020)
ngxson committed 48 days ago
Verified 92bb84f7
server : do not default to multiple slots with speculative decoding (#17017)
ggerganov committed 48 days ago
Verified 13b339bc
mtmd: improve struct initialization (#16981)
ngxson committed 48 days ago
Verified 2f0c2db4
docs: Clarify the endpoint that webui uses (#17001)
openingnow committed 48 days ago
Verified fd2f84f4
model : add openPangu-Embedded (#16941)
Lpzhan931 committed 48 days ago
Verified 9f052478
ggml webgpu: minor set rows optimization (#16810)
reeselevine committed 48 days ago
Verified 03ea0417
sync : ggml
ggerganov committed 48 days ago
cdabeb2c
ggml : fix conv2d_dw SVE path (ggml/1380)
ggerganov committed 48 days ago
852ce518
CUDA: update ops.md (#17005)
mnehete32 committed 48 days ago
Verified 9aa63374
opencl: update doc (#17011)
lhez committed 48 days ago
Verified 5e90233b
refactor: replace sprintf with snprintf for safer string handling in dump functions (#16913)
chraac committed 49 days ago
Verified a5c07dcd
vulkan: remove the need for the dryrun (#16826)
jeffbolznv committed 49 days ago
Verified ad51c0a7
server : do context shift only while generating (#17000)
ggerganov committed 49 days ago
Verified 66d8eccd
readme : update hot topics (#17002)
ggerganov committed 49 days ago
Verified afd35324
ggml-cpu : bicubic interpolation (#16891)
Acly committed 49 days ago
Verified cc98f8d3
ci : apply model label to models (#16994)
CISC committed 49 days ago
Verified d9458343
chore : fix models indent after refactor (#16992)
CISC committed 49 days ago
Verified b164259b
Fix garbled output with REPACK at high thread counts (#16956)
NoahOksuz committed 49 days ago
Verified 1f5accb8
CUDA: avoid mul + bias fusion when doing fusion (#16935)
am17an committed 49 days ago
Verified 2759ccdb
Terms Of Use Privacy Policy Refund Policy Impressum