ngxson/llama.cpp

Pull Requests Commits

jinja: correct stats for tojson and string filters

ngxson committed 170 days ago

3e33f8da

Improve CUDA graph capture (#19754)

gaugarg-nv committed 170 days ago

Verified a0c91e8f

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

crsawyer committed 170 days ago

Verified 07968d53

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

mengshengwu committed 171 days ago

Verified ba3b9c88

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

aldehir committed 171 days ago

Verified 94b0200a

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

taimur-10x committed 171 days ago

Verified b908baf1

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

ddh0 committed 171 days ago

Verified 492bc319

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

jeffbolznv committed 172 days ago

Verified 77d6ae4a

WebUI hide models in router mode (#19374)

crsawyer committed 172 days ago

Verified 10b26ee2

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

jesseposner committed 172 days ago

Verified 3dadc88b

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

abhijitb11 committed 172 days ago

Verified 39e4b1dc

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

yomaytk committed 172 days ago

Verified 11c325c6

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

megemini committed 172 days ago

Verified 237958db

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

0cc4m committed 172 days ago

Verified abb9f3c4

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

ggerganov committed 172 days ago

Verified da348c9d

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

sfallah committed 172 days ago

Verified e6267a93

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

alielfilali01 committed 172 days ago

Verified 2bf318fd

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

JohannesGaessler committed 172 days ago

Verified c78e6822

mtmd : chat : Fix extra \n between text and media marker (#19595)

tdakhran committed 172 days ago

Verified c5897995

webui: Fix Attachments not being included in completion request (#19731)

allozaur committed 172 days ago

Verified 03fd9d3b

model : add tokenizer from LFM2.5-Audio-1.5B (#19687)

tdakhran committed 172 days ago

Verified 8004f3a8

llama : use output_resolve_row() in get_logits_ith/get_embeddings_ith (#19663)

danbev committed 172 days ago

Verified eacb4b67

model : full modern bert support (#18330)

ryan-mangeno committed 172 days ago

Verified c0d04303

llamafile: powerpc: add FP16 MMA path for Q4/Q8 matmul (#19709)

shalinib-ibm committed 172 days ago

Verified 3bb2fcc8

models : dedup qwen35 graphs (#19660)

ggerganov committed 172 days ago

Verified 27326bfc

models : dedup Kimi Linear delta net implementation (#19668)

ymcki committed 172 days ago

Verified ad9f692f

Add Jinja support for "indent" string filter (#19529)

pwilkin committed 173 days ago

Verified 8a709735

ggml webgpu: Fix bug in dispatching large matrix-vector multiplication (#19535)

reeselevine committed 173 days ago

Verified e7f2f95c

server: save generated text for the /slots endpoint (for LLAMA_SERVER_SLOTS_DEBUG=1) (#19622)

matteoserva committed 173 days ago

Verified b55dcdef

model: support GLM-OCR (#19677)

ngxson committed 173 days ago

Verified eeef3cfc

Older