Chenguang Li
|
ca709e427b
CANN: add support for partial RoPE and Vision mode (#17543)
|
1 месяц назад |
Johannes Gäßler
|
0cdce38a97
CUDA: fix FP16 overflow in tile FA kernel (#17875)
|
1 месяц назад |
Aldehir Rojas
|
e39502e74b
llama : add token matching support to llama-grammar (#17816)
|
1 месяц назад |
philip-essential
|
1d2a1ab73d
model : support Rnj-1 (#17811)
|
1 месяц назад |
Sigbjørn Skjæret
|
c8554b66e0
graph : use fill instead of scale_bias in grouped expert selection (#17867)
|
1 месяц назад |
Daniel Bevenius
|
2fa51c19b0
model-conversion : add token ids to prompt token output [no ci] (#17863)
|
1 месяц назад |
Xuan-Son Nguyen
|
951520ddb0
server: delegate result_state creation to server_task (#17835)
|
1 месяц назад |
Neo Zhang
|
68522c678d
ci : support bfloat16 SYCL release package (#17855)
|
1 месяц назад |
Xuan-Son Nguyen
|
f896d2c34f
server: improve speed of speculative decoding (#17808)
|
1 месяц назад |
Piotr Wilkin (ilintar)
|
e4e9c4329c
Make graph_max_nodes vary by ubatch size (#17794)
|
1 месяц назад |
hksdpc255
|
636fc17a37
Fix Kimi-K2 tool-call parsing issues (#17376)
|
1 месяц назад |
Jay Zenith
|
51e0c2d917
cuda : add FILL op support (#17851)
|
1 месяц назад |
Xuan-Son Nguyen
|
37a4f63244
server : add development documentation (#17760)
|
1 месяц назад |
Georgi Gerganov
|
2bc96931d2
server : make cache_reuse configurable per request (#17858)
|
1 месяц назад |
wsbagnsv1
|
5814b4dce1
cuda: optimize SOLVE_TRI using registers and FMAF (#17703)
|
1 месяц назад |
ixgbe
|
79d61896d3
ggml-cpu: add ggml_thread_cpu_relax with Zihintpause support (#17784)
|
1 месяц назад |
Xuan-Son Nguyen
|
4d3726278b
model: add llama 4 scaling for mistral-large (deepseek arch) (#17744)
|
1 месяц назад |
lovedheart
|
08f9d3cc1d
Vulkan: improve mul_mat_vec_iq1_m (#16907)
|
1 месяц назад |
Sigbjørn Skjæret
|
0a540f9abd
ci : add windows-cuda 13.1 release (#17839)
|
1 месяц назад |
Sigbjørn Skjæret
|
22577583a3
common : change --color to accept on/off/auto, default to auto (#17827)
|
1 месяц назад |
Law Po Ying
|
d9e03db1e7
sycl: add missing BF16 conversion support for Intel oneAPI (#17780)
|
1 месяц назад |
Jeff Bolz
|
db97837385
vulkan: perf_logger improvements (#17672)
|
1 месяц назад |
Vishal Singh
|
017761daf5
ggml-zendnn : add ZenDNN backend for AMD CPUs (#17690)
|
1 месяц назад |
Xuan-Son Nguyen
|
c42712b056
server: support multiple generations from one prompt (OAI "n" option) (#17775)
|
1 месяц назад |
Phylliida Dev
|
09c7c50e64
ggml : add circular tiling support to pad, for Vulkan, CUDA, and CPU (used for making seamless textures) (#16985)
|
1 месяц назад |
Johannes Gäßler
|
f334b79494
HIP: fix RDNA3 FP16/BF16 matrix multiplication (#17817)
|
1 месяц назад |
Aleksander Grygier
|
a28e3c7567
webui: Stop generation from chat sidebar (#17806)
|
1 месяц назад |
Aleksander Grygier
|
e31b5c55c3
webui: Fix context available value in Multi-model Router mode (#17804)
|
1 месяц назад |
Aleksander Grygier
|
21f24f27a9
webui: Per-conversation system message with UI displaying, edition & branching (#17275)
|
1 месяц назад |
Sky
|
7b43f55753
ggml : improve error handling for search path existence checks (#17653)
|
1 месяц назад |