ngxson/llama.cpp

Pull Requests Commits

vulkan : fix FA mask load with bounds check (coopmat2) (#17606)

Acly committed 65 days ago

Verified 385c3da5

server: move server-context to its own cpp|h (#17595)

ngxson committed 65 days ago

Verified ab49f094

server: explicitly set the function name in lambda (#17538)

haiyuewa committed 65 days ago

Verified 8c32d9d9

common : fix json schema with '\' in literals (#17307)

i-v-s committed 65 days ago

Verified 0874693b

sycl : support to malloc memory on device more than 4GB, update the doc and script (#17566)

arthw committed 66 days ago

Verified 7d2add51

ggml: replace hwcap with riscv_hwprobe for RVV detection (#17567)

ixgbe committed 66 days ago

Verified f698a79c

Vulkan: MMVQ Integer Dot K-Quant and MUL_MAT_ID support (#16900)

0cc4m committed 66 days ago

Verified 47a268ea

vulkan: improve topk perf for large k, fix overflow in unit tests (#17582)

jeffbolznv committed 66 days ago

Verified 59d8d4e9

gguf-py : fix passing non-native endian tensors (editor-gui and new-metadata) (#17553)

AlekseiNikiforovIBM committed 66 days ago

Verified d82b7a7c

common : move all common_chat_parse_* to chat-parser.cpp. (#17481)

dranger003 committed 66 days ago

Verified 03914c7e

server: fix: /metrics endpoint returning JSON-escaped Prometheus format (#17386)

o7si committed 66 days ago

Verified 3ce7a65c

ggml : add GGML_SCHED_NO_REALLOC option to disable reallocations in ggml_backend_sched (#17276)

slaren committed 67 days ago

Verified e072b205

[MUSA] enable fp16/fast_fp16/bf16_mma on PH1 (#17551)

yeahdongcn committed 67 days ago

Verified c6f7a423

ggml-cuda: add stricter checking for fusion (#17568)

am17an committed 67 days ago

Verified 2e7ef98f

server : add Anthropic Messages API support (#17570)

noname22 committed 67 days ago

Verified ddf9f943

model : Qwen3 Next (#16095)

pwilkin committed 67 days ago

Verified ff55414c

CUDA: no FP16 arithmetic for vector FA kernel (#17558)

JohannesGaessler committed 67 days ago

Verified 73955f7d

vulkan: Implement GGML_OP_TRI (#17503)

jeffbolznv committed 67 days ago

Verified 35cf8887

rpc : cache and reuse compute graphs (#15405)

rgerganov committed 67 days ago

Verified 15d2b46b

HIP: enable mul_mat_f for RDNA4 (#17437)

zhang-hui-yulo committed 67 days ago

Verified 6bca76ff

SOLVE_TRI CUDA kernel for small matrices (#17457)

pwilkin committed 67 days ago

Verified cd0e3a7a

refactor pad_reflect_1d to make the UT case pass (#17204)

NeoZhangJianyu committed 67 days ago

Verified efaaccdd

vulkan: Implement SOLVE_TRI (#17486)

jeffbolznv committed 68 days ago

Verified 4abef75f

arch : add description about LLM_TENSOR_INFOS (#17550)

ggerganov committed 68 days ago

Verified c3861149

models : fix LFM2 tensors (#17548)

ggerganov committed 68 days ago

Verified 6783b11f

cuda : fix UMA detection on discrete GPUs. (#17537)

matt23654 committed 68 days ago

Verified 909072ab

ggml-cpu: aarm64: q4_K repack gemm and gemv implementations (dotprod only) (#17494)

Alcpz committed 68 days ago

Verified cd8370b4

devops: Add build-essential to Ubuntu 26.04 image (#17531)

ericcurtin committed 68 days ago

Verified d21a76ac

gguf-py : skip endian-conversion of MXFP4 data (#17523)

AlekseiNikiforovIBM committed 68 days ago

Verified 4fcd87cf

vulkan : move contiguous checks to device_supports_op (#17490)

Acly committed 68 days ago

Verified b78db3bd

Newer Older