ngxson/llama.cpp

Pull Requests Commits

fix use_kq_norm

ngxson committed 89 days ago

86c8d34b

fix use_kq_norm

ngxson committed 89 days ago

eb601039

half working version

ngxson committed 89 days ago

c540db68

ngxson committed 89 days ago

6f8108c5

convert : add Llama4ForCausalLM

ngxson committed 89 days ago

66fe9ffc

llama-bench: add --n-cpu-moe support (#15952)

jacekpoplawski committed 90 days ago

Verified 8ff20609

ci : use macos-latest for arm64 webgpu build (#16029)

danbev committed 90 days ago

Verified 77475530

ggml : fix padding in timestep embedding kernels (#15932)

danbev committed 90 days ago

Verified 3913f873

ci : upload xcframework artifact from ios-xcode-build job (#16010)

danbev committed 90 days ago

Verified 76888d20

fix: apply clang-format to CUDA macros (#16017)

bugparty committed 90 days ago

Verified f1fbffb5

ci : update macos-latest* jobs to use macos-latest (#15938)

danbev committed 90 days ago

Verified 51abc96b

cmake : Do not install tools on iOS targets (#15903)

ykhrustalev committed 90 days ago

Verified 07808ebb

Add LLaDA-7b-MoE diffusion model (#16003)

am17an committed 90 days ago

Verified 6d758839

CUDA: fix im2col_3d to respect non-contiguous inputs (views) (#15956)

jakekarnes42 committed 91 days ago

Verified 3d4053f7

docker : enable rocWMMA in ROCm images, add gfx1151 (#15997)

slaren committed 91 days ago

Verified dc381aa9

releases : switch to rocWMMA develop branch, add gfx1151 (#15992)

slaren committed 91 days ago

Verified 10d19740

SYCL: Add COUNT_EQUAL operator support (#15991)

yael-works committed 91 days ago

Verified b907255f

llama-run: Fix model download on Windows (#15988)

npopov-vst committed 91 days ago

Verified 28c39da7

CUDA: some micro-optimizations in mmf.cuh for mul_mat_id (#15926)

am17an committed 91 days ago

Verified 10622056

fix KLD percentile output (#15999)

ddh0 committed 91 days ago

Verified a68f31ed

model : add grok-2 support (#15539)

CISC committed 92 days ago

Verified b8e09f08

server : only attempt to enable thinking if using jinja (#15967)

CISC committed 92 days ago

Verified 6c019cb0

metal : remove memory pools (#15966)

ggerganov committed 92 days ago

Verified 9dcd200d

rocm.Dockerfile: added gfx1200,gfx1201 architectures to support AMD Radeon RX 9000 series (#15994)

channeladam committed 92 days ago

Verified 0fa154e3

Vulkan: Clean up mul_mm shader (#15987)

0cc4m committed 92 days ago

Verified 261e6a20

build: fix the build failures of Windows HIP release job (#15984)

lcy0321 committed 92 days ago

Verified a0e13dcb

metal : fix kernel requirements (#15983)

ggerganov committed 92 days ago

Verified a14bd350

rpc : fix regression when --device is used (#15981)

rgerganov committed 92 days ago

Verified 918b26f1

releases : update ROCM, add gfx1200, gfx1201, gfx1151 (#15972)

slaren committed 92 days ago

Verified 9ecb8843

doc : update documentation for --tensor-split (#15980)

rgerganov committed 92 days ago

Verified d1c6f11f

Older