ggerganov/llama.cpp

Pull Requests Commits

llama-bench: add `-d` depth arg (#13096)

thevishalagarwal committed 165 days ago

Verified 1831f538

mtmd : fix glm-edge redundant token count (#13139)

ngxson committed 165 days ago

Verified 4e87962e

context : do not clear output buffer on reserve (#13152)

pockers21 committed 165 days ago

Verified fb0471d1

llama : (mrope) allow using normal 1D position for text token (#13138)

ngxson committed 165 days ago

Verified d2b2031e

clip : refactor set input for cgraph + fix qwen2.5vl input (#13136)

ngxson committed 165 days ago

Verified 5fa9e63b

SYCL: Add all missing unary kernels (#13074)

qnixsynapse committed 165 days ago

Verified a4c340f9

readme : update hot topics (#13150)

ggerganov committed 165 days ago

Verified d0a417f3

common : fix noreturn compile warning (#13151)

ggerganov committed 165 days ago

Verified 43f2b071

llama-chat : fix typo GML --> GLM (#13143)

ngxson committed 165 days ago

Verified e5d6c255

musa: fix typo in cc control (#13144)

yeahdongcn committed 165 days ago

Verified f0dd6a19

CUDA: fix q_nope_absorbed prec for DS 2 Lite f16 (#13137)

JohannesGaessler committed 165 days ago

Verified 69699be4

arg : fix unused variable (#13142)

ngxson committed 165 days ago

Verified 85f36e5e

llama-bench : Add `--override-tensors` arg (#12922)

4onen committed 166 days ago

Verified c0a97b76

llama-chat : fix wrong template in GLM4-0414 (#13140)

matteoserva committed 166 days ago

Verified ced44be3

musa: fix build warning (#13129)

yeahdongcn committed 166 days ago

Verified e291450b

Fixes Qwen2.5VL segfault during inference with https://github.com/ggml-org/llama.cpp/pull/12402 as has_qwen2vl_merger migration was incomplete (#13133)

LostRuins committed 166 days ago

Verified 59e991c2

clip : Add Qwen2.5VL support (#12402)

HimariO committed 166 days ago

Verified ca2bb89e

common : add common_remote_get_content (#13123)

ngxson committed 167 days ago

Verified 2d451c80

clip : improve projector naming (#13118)

ngxson committed 167 days ago

Verified 4753791e

ggml: move fp16/bf16 conversion optimizations to CPU backend + export conversion APIs (#13107)

SongXiaoXi committed 167 days ago

Verified 77d5e9a7

grammar : handle maxItems == 0 in JSON schema (#13117)

rick-github committed 167 days ago

Verified d5fe4e81

llama : fix K-shift with quantized K and BLAS backend (#13113)

slaren committed 168 days ago

Verified 295354ea

Force FP32 compute in GLM4 FFN Down (#13101)

city96 committed 168 days ago

Verified 558a7647

clip : fix pixtral on some GPU backends (#13097)

ngxson committed 168 days ago

Verified edb18b6e

change the reorder tensor from init to execute OP (#13003)

NeoZhangJianyu committed 168 days ago

Verified 514c4560

rpc : do not wait for response when sending RPC_CMD_SET_TENSOR (#12943)

rgerganov committed 168 days ago

Verified 553a5c3a

clip : remove boi/eoi embeddings for GLM-edge model (#13081)

ngxson committed 169 days ago

Verified 13be08da

embeddings : fix batch sizes (#13076)

ggerganov committed 169 days ago

Verified 226251ed

ggml : fix trailing whitespaces (#0)

ggerganov committed 169 days ago

87616f06

ggerganov committed 169 days ago

63b49114

Newer Older