ggerganov/llama.cpp

Pull Requests Commits

llama: add graph build_rope

ngxson committed 13 hours ago

c70e923a

llama: fix kv shifting and add tests

ngxson committed 14 hours ago

1b7f8441

server : realtime: send an error event instead of crashing on exceptions (#31632)

T0mSIlver committed 15 hours ago

Verified 6f1e6543

opencl: improve decode perf with FA (#31633)

shaofeiqi committed 17 hours ago

Verified 9e2a04a4

ui: open actions on right-click for chats, tabs, messages and models (#31545)

allozaur committed 18 hours ago

Verified c6a0ff11

model, mtmd: drama-3 tts support (#31613)

ngxson committed 21 hours ago

Verified 5851a8e5

ui: show text and page thumbnails for PDF attachments, stop endless conversion retries (#31468)

allozaur committed 22 hours ago

Verified efbf6e80

webgpu: return maxBufferSize from get_max_size and avoid whole-buffer bindings in memset/clear (#31348)

yomaytk committed 22 hours ago

Verified b393d73b

sycl : fix docker issue (#31605)

arthw committed 22 hours ago

Verified 00df7c22

gguf-py : validate every tokenizer merge pair, not only the first (#31631)

arxhr007 committed 22 hours ago

Verified 034e2535

common : delete common_get_path_from_env() (#31444)

angt committed 22 hours ago

Verified 52e53a76

llama.cpp : bump version to 0.9.0 (#31646)

ggerganov committed 23 hours ago

Verified cd959bae

ggerganov committed 1 day ago

13ec8f61

ggml : bump version to 0.29.0 (ggml/1680)

ggerganov committed 1 day ago

cca14f64

llama : follow ctx_other devices and flash attention in Gemma4 assistant (#31635)

am17an committed 1 day ago

Verified e853ec9a

ggml : Gate RVV code on the target extension (#31607)

catap committed 1 day ago

Verified feaf8000

common : do not allow `_` in sidecar (#31639)

ggerganov committed 1 day ago

Verified 4c4ca5b1

ggml : fix bf16 (#31609)

EugeoSynthesisThirtyTwo committed 1 day ago

Verified 162e2bb2

opencl: improve q2_0 GEMV for Adreno (#31577)

lhez committed 1 day ago

Verified 6900fb5a

ggml: bound views by the bytes their strides reach (#31584)

ServeurpersoCom committed 1 day ago

Verified bbb02c93

ggml : write host-buffer tensor data directly in gguf writer (#31144)

ggerganov committed 1 day ago

Verified 960b1d51

common : fix -hf tag matching inside longer quant names (#31620)

ggerganov committed 1 day ago

Verified 4b29d026

hexagon: add LIGHTNING_INDEXER support (#31289)

tboinovski1 committed 1 day ago

Verified bee23322

mtmd: accurate support for qwen3-vl video prompt format (#31619)

ngxson committed 1 day ago

Verified 149f5641

model, mtmd: add chatterbox TTS support (#30993)

ngxson committed 1 day ago

Verified b9762796

hexagon: add TOP_K_MASK op (#31552)

tboinovski1 committed 1 day ago

Verified 7071ba37

gemma4: graph optimizations (massive gains on pp for E2B and E4B) (#31356)

max-krasnyansky committed 1 day ago

Verified 6cb3e672

hexagon: faster TOP_K for the backend sampler (#31279)

zhiyuan8 committed 1 day ago

Verified 6b9c0cfc

ui: run local browser tests in installed Chrome with its sandbox on (#31596)

allozaur committed 1 day ago

Verified d1c4e5be

hexagon: add DSV4_HC_PRE and DSV4_HC_POST ops (#30953)

tboinovski1 committed 1 day ago

Verified 20a51f24

Older