ggerganov/llama.cpp

Pull Requests Commits

Merge branch 'master' into xsn/mtmd_custom_min_max_tokens

ngxson committed 38 days ago

79b98dbf

model: add Janus Pro for image understanding (#16906)

ravenouse committed 38 days ago

Verified 6b9a5242

clip : use FA (#16837)

ggerganov committed 38 days ago

Verified 2f966b8e

server : support unified cache across slots (#16736)

ggerganov committed 39 days ago

Verified cd5e3b57

common : move gpt-oss reasoning processing to init params (#16937)

aldehir committed 39 days ago

Verified 87c9efc3

docs: remove llama_sampler_accept reference in sampling sample usage (#16920)

alundb committed 39 days ago

Verified 76af40aa

CUDA: add FLOOR, CEIL, ROUND, TRUNC unary ops (#16917)

mnehete32 committed 39 days ago

Verified 7db35a79

devops: fix failing s390x docker build (#16918)

taronaeo committed 39 days ago

Verified a864132b

ggml: add s390x cpu-feats (#16774)

taronaeo committed 39 days ago

Verified d38d9f08

scripts : add script to bench models (#16894)

ggerganov committed 39 days ago

Verified 7fd205a8

webui: auto-refresh /props on inference start to resync model metadata (#16784)

ServeurpersoCom committed 39 days ago

Verified 2f68ce7c

mtmd: add --image-min/max-tokens

ngxson committed 39 days ago

70cc3300

webui: add HTML/JS preview support to MarkdownContent with sandboxed iframe (#16757)

ServeurpersoCom committed 40 days ago

Verified e4a71599

vendor : update cpp-httplib to 0.27.0 (#16846)

angt committed 40 days ago

Verified dd5e8cab

mtmd: refactor preprocessing + support max/min pixels (#16878)

ngxson committed 40 days ago

Verified cf659bbb

Add a setting to display message generation statistics (#16901)

allozaur committed 40 days ago

Verified d8b860a2

webui: recognize AsciiDoc files as valid text files (#16850)

jhradilek committed 40 days ago

Verified 1ae74882

common : allow --system-prompt-file for diffusion-cli (#16903)

CISC committed 40 days ago

Verified 961660b8

codeowners : update after refactor (#16905)

CISC committed 40 days ago

Verified 74fef412

vulkan: Fix multi_add invalid descriptor usage (#16899)

jeffbolznv committed 40 days ago

Verified 5d8bb900

vulkan: fuse mul_mat+add and mul_mat_id+add_id (#16868)

jeffbolznv committed 40 days ago

Verified 2e76e013

CUDA: Remove unneded bias/gate dims in fused mmvq (#16858)

ORippler committed 40 days ago

Verified d3dc9dd8

refactor : llama-model.cpp (#16252)

pwilkin committed 40 days ago

Verified bea04522

model : Minimax M2 (#16831)

pwilkin committed 40 days ago

Verified 0de0a015

model : add Granite Hybrid nano types (#16896)

giuseppe committed 40 days ago

Verified e58d5856

CUDA: Volta tensor core support for MMF (#16843)

JohannesGaessler committed 41 days ago

Verified 31c511a9

ggerganov committed 41 days ago

6d39015a

CUDA: add expert reduce kernel (#16857)

am17an committed 41 days ago

Verified 4146d6a1

batch : fix consistency checks for the input positions (#16890)

ggerganov committed 41 days ago

Verified 8da3c0e2

server : don't print user inputs to console (#16871)

ggerganov committed 41 days ago

Verified c22473b5

Older