ngxson/llama.cpp

Pull Requests Commits

ngxson committed 54 days ago

f4e70a48

mtmd: add pos_0 to mtmd_image_tokens_get_decoder_pos

ngxson committed 54 days ago

3b2a68ab

android : libcommon -> libllama-common (#22076)

CISC committed 54 days ago

Verified 23b8cc49

ggml-backend-meta: add multi-segment read support in get_tensor (#22063)

ssam18 committed 54 days ago

Verified 59accc88

ci : free disk space for rocm release (#22012)

CISC committed 54 days ago

Verified 83d58e02

convert : fix (ignore for now) typings errors (#22002)

CISC committed 54 days ago

Verified 89a5474f

llama: fit ctx size for CPU only (#21568)

JohannesGaessler committed 54 days ago

Verified fd1c0ec3

ggml-webgpu: fix compiler warnings and refactor FlashAttention encoding (#21052)

reeselevine committed 55 days ago

Verified 45cac7ca

CUDA: use LRU based eviction for cuda graphs (#21611)

am17an committed 55 days ago

Verified b94050e8

ci : add android arm64 build and release (#21647)

ykhrustalev committed 55 days ago

Verified a279d0f0

mtmd: add missing struct tag (#22023)

65a committed 55 days ago

Verified 268d61e1

libs : rename libcommon -> libllama-common (#21936)

ggerganov committed 55 days ago

Verified 6990e2f1

model : Gemma4 model type detection (#22027)

EZForever committed 55 days ago

Verified fcc75087

opencl: refactor q8_0 set_tensor and mul_mat host side dispatch for Adreno (#21938)

lhez committed 56 days ago

Verified 5e6c0e18

cli : use get_media_marker (#22017)

CISC committed 56 days ago

Verified 30dce2cf

cmake: use glob to collect src/models sources (#22005)

ngxson committed 56 days ago

Verified 089dd41f

hexagon: optimize HMX matmul operations (#21071)

chraac committed 56 days ago

Verified 85dde8dc

model: using single llm_build per arch (#21970)

ngxson committed 56 days ago

Verified 4fbdabdc

opencl: add q5_K gemm and gemv kernels for Adreno (#21595)

shaofeiqi committed 56 days ago

Verified e45dbdec

server: tests: fetch random media marker via /apply-template (#21962) (#21980)

ServeurpersoCom committed 56 days ago

Verified 4adac43f

model : refactor QKV into common build_qkv and create_tensor_qkv helpers (#21245)

JoursBleu committed 56 days ago

Verified 9db77a02

model : support NVFP4 tensors for Gemma4 (#21971)

CISC committed 56 days ago

Verified f772f6e4

codeowners: add team member comments (#21714)

0cc4m committed 56 days ago

Verified b572d1ec

Convert: Fix NemotronH Config Parsing (#21664)

anavp-nvidia committed 56 days ago

Verified 03b3d077

ggml: add graph_reused (#21764)

am17an committed 56 days ago

Verified 3f7c29d3

metal: Implement ROLL op (#21946)

kushagharahi committed 56 days ago

Verified ae2d3489

ggml-cpu: add 128-bit RVV implementation for Quantization Vector Dot (#20633)

rehan-10xengineer committed 56 days ago

Verified 1e796eb4

ggml : implemented simd_gemm kernel for riscv vector extension (#20627)

rehan-10xengineer committed 56 days ago

Verified 56375365

devops : added spirv-headers to nix (#21965)

yuannan committed 56 days ago

Verified 90fb96a7

ggml-webgpu: compute pass batching and removing profiling overhead (#21873)

reeselevine committed 56 days ago

Verified 82677a6e

Older