 Xuan-Son Nguyen
|
3464bdac37
llama: fix ASAN error with M-RoPE (#16848)
|
пре 2 месеци |
|
Xuan-Son Nguyen
|
e3af5563bd
llama: store mrope data in KV cell (#16825)
|
пре 2 месеци |
 Jeff Bolz
|
10fcc41290
vulkan: Update topk_moe fusion to handle gpt's late softmax (#16656)
|
пре 2 месеци |
 Ruben Ortlam
|
bcf5bda6f5
Vulkan MMQ Integer Dot Refactor and K-Quant support (#16536)
|
пре 2 месеци |
 Max Krasnyansky
|
3eb2be1ca5
Hexagon Op queue & dispatch optimizations (#16820)
|
пре 2 месеци |
 Aman Gupta
|
e41bcce8f0
CUDA: use fastdiv in set-rows (#16834)
|
пре 2 месеци |
 Sigbjørn Skjæret
|
144a4ce824
vendor : sync minja (#16500)
|
пре 2 месеци |
|
Jeff Bolz
|
f549b0007d
vulkan: Call ggml_vk_buffer_write_2d from ggml_vk_buffer_copy (#16793)
|
пре 2 месеци |
|
Aman Gupta
|
9a3ea685b9
CUDA: Fix bug in topk-moe for gpt-oss (#16821)
|
пре 2 месеци |
 YaelLogic
|
338074c383
sycl: add RMS_NORM_BACK operation support (#16808)
|
пре 2 месеци |
 YaelGitAccount
|
851553ea6b
cuda: add SET operation support (#16804)
|
пре 2 месеци |
 Georgi Gerganov
|
85a7d8677b
memory : remove KV cache size padding (#16812)
|
пре 2 месеци |
|
Georgi Gerganov
|
a8ca18b4b8
llama-bench : clarify benchmarked parts of the computation (#16823)
|
пре 2 месеци |
 l3utterfly
|
8284efc35c
initialise buffer.device in ggml_hexagon_session (#16816)
|
пре 2 месеци |
 Sam Malayek
|
1c1409e131
embedding: add raw option for --embd-output-format (#16541)
|
пре 2 месеци |
 Johannes Gäßler
|
7a0e900e36
llama: consistent ctx <-> buf order for KV cache (#16746)
|
пре 2 месеци |
 Aldehir Rojas
|
280d97be96
grammar : support array references in json schema (#16792)
|
пре 2 месеци |
 Chenguang Li
|
3479efd112
CANN: Improve device ID handling and aclnnArange checks (#16752)
|
пре 2 месеци |
|
Aman Gupta
|
463bbf20bf
CUDA: add unused vars to mmvf and mmvq (#16807)
|
пре 2 месеци |
 tamarPal
|
ad8d36beff
sycl: add SSM_CONV operation support (#16800)
|
пре 2 месеци |
 Yuri Khrustalev
|
c053e18a66
chat: Add LFM2 tool handling (#16763)
|
пре 2 месеци |
|
Xuan-Son Nguyen
|
e1ab084803
mtmd : fix idefics3 preprocessing (#16806)
|
пре 2 месеци |
 Diego Devesa
|
5a4ff43e7d
llama : disable pipeline parallelism if compute buffer allocation fails (#16748)
|
пре 2 месеци |
 Acly
|
10640e31aa
ggml : fix interpolate with align-corners and ne=1 (#16700)
|
пре 2 месеци |
|
Johannes Gäßler
|
80d28f104c
HIP: fix AMDGPU_TARGETS, update documentation (#16803)
|
пре 2 месеци |
|
Xuan-Son Nguyen
|
c55d53acec
model : add LightOnOCR-1B model (#16764)
|
пре 2 месеци |
|
Johannes Gäßler
|
945501f5ea
llama: fix leaked buffers for mmap + split files (#16765)
|
пре 2 месеци |
|
Aman Gupta
|
75cbdd3fce
test-backend-ops: print failed tests at the end (#16785)
|
пре 2 месеци |
|
tamarPal
|
2b9bd9bf4e
sycl: add ROLL operation support (#16665)
|
пре 2 месеци |
 shani-f
|
59fc1ec8e8
sycl: add REPEAT_BACK operation support (#16734)
|
пре 2 месеци |
