| .. |
|
amx
|
a23b9bdbd3
ggml : fix unaligned access in AMX code (#16315)
|
3 months ago |
|
arch
|
df70bedda7
arm64: add i8mm route with SVE ggml_vec_dot_q4_K_q8_K and ggml_vec_dot_q6_K_… (#15277)
|
2 months ago |
|
cmake
|
ae8de6d50a
ggml : build backends as libraries (#10256)
|
1 year ago |
|
kleidiai
|
8c583242ad
kleidiai: add optimized per-channel kernels for Q8_0 (#16993)
|
2 months ago |
|
llamafile
|
a6a58d6478
llamafile: PowerPC Sgemm Optimization (#15558)
|
5 months ago |
|
spacemit
|
342c728d03
ggml : fix SpaceMit IME array out-of-bounds in task assignment (#16629)
|
3 months ago |
|
CMakeLists.txt
|
8c583242ad
kleidiai: add optimized per-channel kernels for Q8_0 (#16993)
|
2 months ago |
|
arch-fallback.h
|
9b26511857
ggml-cpu: implement MXFP4 SIMD for s390x (#16193)
|
4 months ago |
|
binary-ops.cpp
|
a62d7fa7a9
cpu: de-duplicate some of the operators and refactor (ggml/1144)
|
10 months ago |
|
binary-ops.h
|
a62d7fa7a9
cpu: de-duplicate some of the operators and refactor (ggml/1144)
|
10 months ago |
|
common.h
|
0dd58b6877
ggml : refactor forward_dup for cpu backend (#16062)
|
4 months ago |
|
ggml-cpu-impl.h
|
fcfce040e8
ggml : LoongArch fixes (#16958)
|
2 months ago |
|
ggml-cpu.c
|
395e286bc9
cpu: skip NOPs to avoid barriers (#17133)
|
2 months ago |
|
ggml-cpu.cpp
|
b77e6c18e1
ggml: riscv: add riscv spacemit backend (#15288)
|
3 months ago |
|
hbm.cpp
|
f470bc36be
ggml-cpu : split arch-specific implementations (#13892)
|
7 months ago |
|
hbm.h
|
f470bc36be
ggml-cpu : split arch-specific implementations (#13892)
|
7 months ago |
|
ops.cpp
|
395e286bc9
cpu: skip NOPs to avoid barriers (#17133)
|
2 months ago |
|
ops.h
|
395e286bc9
cpu: skip NOPs to avoid barriers (#17133)
|
2 months ago |
|
quants.c
|
fd1234cb46
llama : add gpt-oss (#15091)
|
5 months ago |
|
quants.h
|
fd1234cb46
llama : add gpt-oss (#15091)
|
5 months ago |
|
repack.cpp
|
1f5accb8d0
Fix garbled output with REPACK at high thread counts (#16956)
|
2 months ago |
|
repack.h
|
00f35d509e
ggml : repack block_iq4_nlx8 (#14904)
|
5 months ago |
|
simd-mappings.h
|
fcfce040e8
ggml : LoongArch fixes (#16958)
|
2 months ago |
|
traits.cpp
|
0d8831543c
ggml : fix fallback to CPU for ununsupported ops (#15118)
|
5 months ago |
|
traits.h
|
0d8831543c
ggml : fix fallback to CPU for ununsupported ops (#15118)
|
5 months ago |
|
unary-ops.cpp
|
466c1911ab
cpu : add FLOOR, CEIL, ROUND and TRUNC unary operators (#16083)
|
3 months ago |
|
unary-ops.h
|
466c1911ab
cpu : add FLOOR, CEIL, ROUND and TRUNC unary operators (#16083)
|
3 months ago |
|
vec.cpp
|
c515fc5771
ggml : fix scalar path for computing norm (#16558)
|
3 months ago |
|
vec.h
|
a2e0088d92
Revert "ggml : Leverage the existing GGML_F32_VEC helpers to vectorize ggml_v…" (#16723)
|
3 months ago |
Charles Xu
8c583242ad
kleidiai: add optimized per-channel kernels for Q8_0 (#16993)