 Georgi Gerganov
|
d774ab3acc
metal : adjust support conditions for norm operators (#11671)
|
11 months ago |
 Johannes Gäßler
|
fa62da9b2d
CUDA: support for mat. mul. with ne03 != ne13 (#11656)
|
11 months ago |
 SAMI
|
1ec208083c
llava: add quantization for the visual projector LLAVA, Qwen2VL (#11644)
|
11 months ago |
 Olivier Chafik
|
9f4cc8f8d3
`sync`: minja (#11641)
|
11 months ago |
|
Johannes Gäßler
|
fd08255d0d
CUDA: non-contiguous (RMS) norm support (#11659)
|
11 months ago |
 fxzjshm
|
3ec9fd4b77
HIP: force max threads per block to be 1024 (#11621)
|
11 months ago |
 Xuan-Son Nguyen
|
3962fc1a79
server : add try..catch to places not covered by set_exception_handler (#11620)
|
11 months ago |
 Radoslav Gerganov
|
1bef571f6a
arg : list RPC devices first when using --list-devices (#11655)
|
11 months ago |
|
Olivier Chafik
|
db288b60cb
`tool-call`: command r7b fix for normal responses (#11608)
|
11 months ago |
 Shelby Jenkins
|
106045e7bb
readme : add llm_client Rust crate to readme bindings (#11628)
|
11 months ago |
 Jhen-Jie Hong
|
f117d84b48
swift : fix llama-vocab api usage (#11645)
|
11 months ago |
|
Jhen-Jie Hong
|
534c46b53c
metal : use residency set for other platforms (#11648)
|
11 months ago |
|
Georgi Gerganov
|
387a1598ca
authors : update
|
11 months ago |
|
Georgi Gerganov
|
7c9e0ca520
sync : ggml
|
11 months ago |
 Christian Kastner
|
8f8290ada9
cmake: Add ability to pass in GGML_BUILD_NUMBER (ggml/1096)
|
11 months ago |
|
Georgi Gerganov
|
b34aedd558
ci : do not stale-close roadmap issues
|
11 months ago |
|
Olivier Chafik
|
cde3833239
`tool-call`: allow `--chat-template chatml` w/ `--jinja`, default to chatml upon parsing issue, avoid double bos (#11616)
|
11 months ago |
|
Xuan-Son Nguyen
|
b3451785ac
server : (webui) revert hacky solution from #11626 (#11634)
|
11 months ago |
 Woof Dog
|
1d1e6a90bc
server : (webui) allow typing and submitting during llm response (#11626)
|
11 months ago |
 Daniel Bevenius
|
5598f475be
server : remove CPPHTTPLIB_NO_EXCEPTIONS define (#11622)
|
11 months ago |
|
Georgi Gerganov
|
8ec05832fa
sync : ggml
|
11 months ago |
|
Johannes Gäßler
|
21c84b5d2d
CUDA: fix Volta FlashAttention logic (#11615)
|
11 months ago |
 mashdragon
|
d92cb67e37
server : (webui) Fix Shift+Enter handling (#11609)
|
11 months ago |
|
Johannes Gäßler
|
6eecde3cc8
HIP: fix flash_attn_stream_k_fixup warning (#11604)
|
11 months ago |
 uvos
|
396856b400
CUDA/HIP: add support for selectable warp size to mmv (#11519)
|
11 months ago |
|
uvos
|
4d0598e144
HIP: add GGML_CUDA_CC_IS_* for amd familys as increasing cc archtectures for amd gpus are not supersets of eatch other (#11601)
|
11 months ago |
|
Olivier Chafik
|
90f9b88afb
nit: more informative crash when grammar sampler fails (#11593)
|
11 months ago |
|
Johannes Gäßler
|
864a0b67a6
CUDA: use mma PTX instructions for FlashAttention (#11583)
|
11 months ago |
 Eric Curtin
|
84ec8a58f7
Name colors (#11573)
|
11 months ago |
|
Olivier Chafik
|
bfcce4d693
`tool-call`: support Command R7B (+ return tool_plan "thoughts" in API) (#11585)
|
11 months ago |
