 Georgi Gerganov
|
f0678c5ff4
ggml : fix I8MM Q4_1 scaling factor conversion (#10562)
|
hai 1 ano |
 Shupei Fan
|
4b3242bbea
ggml-cpu: fix typo in gemv/gemm iq4_nl_4_4 (#10580)
|
hai 1 ano |
 Alberto Cabrera Pérez
|
0f77aae560
sycl : offload of get_rows set to 0 (#10432)
|
hai 1 ano |
|
Alberto Cabrera Pérez
|
266b8519ee
sycl : Reroute permuted mul_mats through oneMKL (#10408)
|
hai 1 ano |
 Chenguang Li
|
938f608742
CANN: RoPE operator optimization (#10563)
|
hai 1 ano |
 Jeff Bolz
|
f095a649ec
vulkan: get the first command buffer submitted sooner (#10499)
|
hai 1 ano |
 Ting Lou
|
678d7994f4
llava: return false instead of exit (#10546)
|
hai 1 ano |
|
Georgi Gerganov
|
dc22344088
ggml : remove redundant copyright notice + update authors
|
hai 1 ano |
|
Georgi Gerganov
|
4c0a95b107
llama : add missing model types
|
hai 1 ano |
 Xuan Son Nguyen
|
6c59567689
server : (tests) don't use thread for capturing stdout/stderr, bump openai client library (#10568)
|
hai 1 ano |
 Johannes Gäßler
|
890719311b
common: fix warning message when no GPU found (#10564)
|
hai 1 ano |
 Random Fly
|
7281cf13ad
docs: fix outdated usage of llama-simple (#10565)
|
hai 1 ano |
 Diego Devesa
|
e90688edd0
ci : fix tag name in cuda and hip releases (#10566)
|
hai 1 ano |
|
Georgi Gerganov
|
76b27d29c2
ggml : fix row condition for i8mm kernels (#10561)
|
hai 1 ano |
|
Georgi Gerganov
|
eea986f215
cmake : fix ARM feature detection (#10543)
|
hai 1 ano |
|
Shupei Fan
|
c202cef168
ggml-cpu: support IQ4_NL_4_4 by runtime repack (#10541)
|
hai 1 ano |
 Sergio López
|
2025fa67e9
kompute : improve backend to pass test_backend_ops (#10542)
|
hai 1 ano |
 Ruixin Huang
|
c6bc73951e
CANN: Update cann.md to display correctly in CLion (#10538)
|
hai 1 ano |
 leo-pony
|
605fa66c50
CANN: Fix SOC_TYPE compile bug (#10519)
|
hai 1 ano |
|
Chenguang Li
|
b7420131bf
CANN: ROPE operator optimization (#10540)
|
hai 1 ano |
|
Xuan Son Nguyen
|
9f912511bc
common : fix duplicated file name with hf_repo and hf_file (#10550)
|
hai 1 ano |
 uvos
|
3ad5451f3b
Add some minimal optimizations for CDNA (#10498)
|
hai 1 ano |
|
Diego Devesa
|
46c69e0e75
ci : faster CUDA toolkit installation method and use ccache (#10537)
|
hai 1 ano |
|
Georgi Gerganov
|
9e2301f4a4
metal : fix group_norm support condition (#0)
|
hai 1 ano |
|
Georgi Gerganov
|
fee824a1a1
sync : ggml
|
hai 1 ano |
 Frankie Robertson
|
9150f8fef9
Do not include arm_neon.h when compiling CUDA code (ggml/1028)
|
hai 1 ano |
|
Jeff Bolz
|
c31ed2abfc
vulkan: define all quant data structures in types.comp (#10440)
|
hai 1 ano |
|
Jeff Bolz
|
5b3466bedf
vulkan: Handle GPUs with less shared memory (#10468)
|
hai 1 ano |
|
Jeff Bolz
|
249a7902ec
vulkan: further optimize q5_k mul_mat_vec (#10479)
|
hai 1 ano |
|
Jeff Bolz
|
71a64989a5
vulkan: skip integer div/mod in get_offsets for batch_idx==0 (#10506)
|
hai 1 ano |
