ochafik/llama.cpp

Pull Requests Commits

ggml-cpu: arm64: q5_K repack gemm and gemv (and generic) implementations (dotprod) (#19356)

Alcpz committed 245 days ago

Verified bc160d35

llama : remove write/read of output ids/logits/embeddings (#18862)

danbev committed 246 days ago

Verified 2b6dfe82

cli : provide model with text filename (#19783)

CISC committed 246 days ago

Verified e8e26169

jinja: correct stats for tojson and string filters (#19785)

ngxson committed 246 days ago

Verified 5452d736

common : fix improper trimming in XML parser on complete message (#19805)

aldehir committed 246 days ago

Verified ed483789

Fix wrong cli-argument in documentation (#19804)

Menkalian committed 246 days ago

Verified cacc371f

model : add Kanana-2 model support (#19803)

HelloKS committed 246 days ago

Verified ae2368e7

ci : fix rocm archive name [no ci] (#19808)

CISC committed 246 days ago

Verified 9f0684f0

server : merge contiguous Responses input items into a single assistant message (#19773)

aldehir committed 246 days ago

Verified 34ec1c3f

ci : fix rocm release path [no ci] (#19784)

CISC committed 247 days ago

Verified e877ad8b

Update ROCm docker container to 7.2 release (#19418)

superm1 committed 247 days ago

Verified 35715657

Add a build target to generate ROCm artifacts using ROCm 7.2 (#19433)

superm1 committed 247 days ago

Verified f75c4e8b

vendor : update cpp-httplib to 0.33.1 (#19778)

angt committed 247 days ago

Verified 99156f3a

Improve CUDA graph capture (#19754)

gaugarg-nv committed 247 days ago

Verified a0c91e8f

fix: UI single model selection in router mode (#19767)

crsawyer committed 247 days ago

Verified 07968d53

hexagon : fix build release (#19444) (#19587)

mengshengwu committed 248 days ago

Verified ba3b9c88

common : merge qwen3-coder and nemotron nano 3 parsers (#19765)

aldehir committed 248 days ago

Verified 94b0200a

ggml-cpu: add RVV vec dot kernels for quantization types (#18784)

taimur-10x committed 248 days ago

Verified b908baf1

quantize : add --dry-run option (#19526)

ddh0 committed 248 days ago

Verified 492bc319

test: mul_mat tests with huge batch size (#19519)

jeffbolznv committed 249 days ago

Verified 77d6ae4a

WebUI hide models in router mode (#19374)

crsawyer committed 249 days ago

Verified 10b26ee2

common : fix Step-3.5-Flash format detection and thinking support (#19635)

jesseposner committed 249 days ago

Verified 3dadc88b

common : fix gpt-oss Jinja error when assistant message has both content and thinking with tool calls (#19704)

abhijitb11 committed 249 days ago

Verified 39e4b1dc

ggml-webgpu: Add unary op (SQR, SQRT, SIN, COS) support. (#19700)

yomaytk committed 249 days ago

Verified 11c325c6

model: Add PaddleOCR-VL model support (#18825)

megemini committed 249 days ago

Verified 237958db

vulkan: fix MMQ shader push constants and multi-dispatch (#19732)

0cc4m committed 249 days ago

Verified abb9f3c4

models : fix qwen3.5 beta/gate shapes (#19730)

ggerganov committed 249 days ago

Verified da348c9d

mtmd: build_attn modified, flash_attn on/off via ctx_params (#19729)

sfallah committed 249 days ago

Verified e6267a93

model : add JAIS-2 architecture support (#19488)

alielfilali01 committed 249 days ago

Verified 2bf318fd

CUDA: fix kernel selection logic for tile FA (#19686)

JohannesGaessler committed 249 days ago

Verified c78e6822

Newer Older