ggerganov/llama.cpp

Pull Requests Commits

grammar : parse \- in char classes as literal hyphen (#27591)

NIXKnight committed 7 days ago

Verified eb25b726

sycl : mark tq2_0 as not supported (#27660)

arthw committed 7 days ago

Verified 814d84bc

webgpu : fix handling of infinity values during ARGSORT and TOP_K (#27538)

fairydreaming committed 7 days ago

Verified 5ea87dda

metal : per-device tuned (Q, NE) for flash-attn vec (#26570)

forforever73 committed 8 days ago

Verified f280b269

metal: per-op source split + parallel compile (#26561)

forforever73 committed 8 days ago

Verified b615f5b4

misc : read repetition_penalty from generation_config.json (#27659)

tdakhran committed 8 days ago

Verified b3c3b96a

tests : disable DOTS3NOTE arch test for WebGPU (#27654)

fairydreaming committed 8 days ago

Verified 75844307

convert: fix GLM regression in index_tensors (#27655)

jacekpoplawski committed 8 days ago

Verified 71cc86fa

ggml : shorten virtual device naming in CUDA and Metal (#27608)

ggerganov committed 8 days ago

Verified a14dba68

webgpu : reorder includes since V that appears in common_decls.tmpl may be defined as K in flash_attn_decls.tmpl if KV_OVERLAP (#27545)

fairydreaming committed 8 days ago

Verified c1c766da

mtmd: video: fix moov atom at the end of file (#27596)

ngxson committed 8 days ago

Verified 160c6b0b

ci : apply ccache-clear with older/min/dry-run to all ccache jobs (#27602)

ggerganov committed 8 days ago

Verified 985b1491

ggml : fix ggml_clamp (#27644)

ggerganov committed 8 days ago

Verified 6036c635

mamba2 : Flatten in/out projections to dispatch GEMM instead of GEMV (#27513)

pskrunner14 committed 8 days ago

Verified a130532a

Deepseek 4: `-sm tensor` (#26490)

am17an committed 8 days ago

Verified bf0a29cc

model : support MTP in GLM-4.5-Air (#26534)

jacekpoplawski committed 9 days ago

Verified c060ca97

readme : update links (#27617)

ggerganov committed 9 days ago

Verified ccc8fd2b

fix: Change chat tabs nav shortcuts (#27609)

allozaur committed 9 days ago

Verified d05f8956

test : fix multi-GPU server tests (#27614)

ggerganov committed 9 days ago

Verified 8d9af256

test: move tools/parser to tests (#27548)

ngxson committed 9 days ago

Verified 4a08fa29

mtmd: use pillow-accurate algo, correct resize_algo for all models (#27594)

ngxson committed 9 days ago

Verified 56db501e

ci : add test-llama-archs tensor split for Metal (#27598)

ggerganov committed 9 days ago

Verified 95b8e33e

contrib : recommend waiting for CI before merging (#27603)

nikwen committed 9 days ago

Verified a278dcef

server : add LLAMA_SERVER_SLOTS_N_DIFF (#27600)

ggerganov committed 9 days ago

Verified e8eed452

common : skip device_info loop if it's not going to be printed (#26692)

wolfpld committed 9 days ago

Verified ba8e0edd

DeepseekV4: fix rollback with multi-seq (#26756)

am17an committed 9 days ago

Verified b0539c43

[Tensor parallel] Fix meta tensor split state propagation (#27574)

gaugarg-nv committed 9 days ago

Verified d3371929

ui: Chat Conversation Tabbed navigation (#27263)

allozaur committed 9 days ago

Verified 8144f319

vendor : update subprocess.h (#27409)

cabelo committed 9 days ago

Verified 6657ded4

cuda : add POOL_1D support (#27573)

amankarki151 committed 9 days ago

Verified 29ea9412

Newer Older