ggerganov/llama.cpp

Pull Requests Commits

ops : update BLAS

ggerganov committed 1 year ago

Verified e9f7e7cc

ops : update Metal (#14912)

ggerganov committed 1 year ago

Verified c35f9eaf

ggerganov committed 1 year ago

1f45f289

cmake : Indent ggml-config.cmake (ggml/1310)

dg0yt committed 1 year ago

613c5095

quantize : update README.md (#14905)

EAddario committed 1 year ago

Verified 7f975995

vulkan: add ops docs (#14900)

0cc4m committed 1 year ago

Verified bf78f543

SYCL: add ops doc (#14901)

qnixsynapse committed 1 year ago

Verified bbfc8492

llama : clarify comment about pp and tg graphs [no ci] (#14895)

danbev committed 1 year ago

Verified ca0ef2dd

vulkan : add fp16 support for the conv_2d kernel (#14872)

Green-Sky committed 1 year ago

Verified 89d10295

vulkan: skip empty set_rows to avoid invalid API usage (#14860)

jeffbolznv committed 1 year ago

Verified f1a4e72d

model : make rope_yarn_log_mul optional for deepseek2 (#14896)

gabriellarson committed 1 year ago

Verified 4762ad73

llama : fix kq_scale for the attention layers of PLaMo2 (#14892)

mitmul committed 1 year ago

Verified 1dc9614e

Docs: add instructions for adding backends (#14889)

am17an committed 1 year ago

Verified 446595b9

HIP: Enable Matrix cores for MMQ Kernels, Enable stream-K for CDNA 3 (#14624)

deepsek committed 1 year ago

Verified 66906cd8

CANN: Implement GLU ops (#14884)

hipudding committed 1 year ago

Verified 11dd5a44

musa: fix build warnings (unused variable) (#14869)

yeahdongcn committed 1 year ago

Verified 9b8f3c6c

ggml-cpu : disable GGML_NNPA by default due to instability (#14880)

taronaeo committed 1 year ago

Verified c7f3169c

metal: SSM_SCAN performance (#14743)

gabe-l-hart committed 1 year ago

Verified 793c0d7f

opencl: add fused `rms_norm_mul` (#14841)

lhez committed 1 year ago

Verified ce111d39

docs : update HOWTO‑add‑model.md for ModelBase and new model classes (#14874)

wooksong committed 1 year ago

Verified e7fecba9

ggml : remove invalid portPos specifiers from dot files (#14838)

ORippler committed 1 year ago

Verified e2b7621e

context : restore preemptive sched reset when LLAMA_SET_ROWS=0 (#14870)

ggerganov committed 1 year ago

Verified c1dbea75

mtmd : fix 32-bit narrowing issue in export-lora and mtmd clip (#14503)

kiwi142857 committed 1 year ago

Verified 749e0d27

rpc : check for null buffers in get/set/copy tensor endpoints (#14868)

struct committed 1 year ago

Verified 64bf1c37

sched : fix multiple evaluations of the same graph with pipeline parallelism (#14855)

slaren committed 1 year ago

Verified c12bbde3

musa: upgrade musa sdk to rc4.2.0 (#14498)

yeahdongcn committed 1 year ago

Verified 3f4fc97f

ggerganov committed 1 year ago

2df255da

cmake : fix usage issues (ggml/1257)

dg0yt committed 1 year ago

60f816a7

ggml-cpu : remove stdlib include from repack.cpp (ggml/1276)

danbev committed 1 year ago

5592f278

context : perform output reorder lazily upon access after sync (#14853)

ggerganov committed 1 year ago

Verified e4868d16

Older