ikawrakow/ik_llama.cpp
1
0
Fork 0
mirror of https://github.com/ikawrakow/ik_llama.cpp.git synced 2026-01-26 17:20:01 +00:00
Code Issues Packages Projects Releases Wiki Activity

correct typo (#876)

Browse Source
This commit is contained in:
Nexes the Elder
2025-10-28 18:01:45 +01:00
committed by GitHub
parent e68dabc242
commit d50c2490fc
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
ggml/src/ggml-cuda/cpy.cu
View File

@@ -336,7 +336,7 @@ static void ggml_cpy_f32_q6_0_cuda(
const int nb03, const int ne10, const int ne11, const int ne12, const int nb10, const int nb11, const int nb12, const int nb13, cudaStream_t stream, char ** cdst_indirect, int & graph_cpynode_index) { const int nb03, const int ne10, const int ne11, const int ne12, const int nb10, const int nb11, const int nb12, const int nb13, cudaStream_t stream, char ** cdst_indirect, int & graph_cpynode_index) {
GGML_ASSERT(ne % QK6_0 == 0); GGML_ASSERT(ne % QK6_0 == 0);
const int num_blocks = ne / QK5_1; const int num_blocks = ne / QK6_0;
cpy_f32_q<cpy_blck_f32_q6_0, QK6_0><<<num_blocks, 1, 0, stream>>> cpy_f32_q<cpy_blck_f32_q6_0, QK6_0><<<num_blocks, 1, 0, stream>>>
(cx, cdst, ne, ne00, ne01, ne02, nb00, nb01, nb02, nb03, ne10, ne11, ne12, nb10, nb11, nb12, nb13, cdst_indirect, graph_cpynode_index++); (cx, cdst, ne, ne00, ne01, ne02, nb00, nb01, nb02, nb03, ne10, ne11, ne12, nb10, nb11, nb12, nb13, cdst_indirect, graph_cpynode_index++);
} }
@@ -590,7 +590,7 @@ void* ggml_cuda_cpy_fn(const ggml_tensor * src0, ggml_tensor * src1) {
} else if (src0->type == GGML_TYPE_F32 && src1->type == GGML_TYPE_Q6_0) { } else if (src0->type == GGML_TYPE_F32 && src1->type == GGML_TYPE_Q6_0) {
return (void*) cpy_f32_q<cpy_blck_f32_q6_0, QK6_0>; return (void*) cpy_f32_q<cpy_blck_f32_q6_0, QK6_0>;
} else if (src0->type == GGML_TYPE_Q6_0 && src1->type == GGML_TYPE_F32) { } else if (src0->type == GGML_TYPE_Q6_0 && src1->type == GGML_TYPE_F32) {
return (void*) cpy_q_f32<cpy_blck_q_f32<dequantize_q6_0, QK6_0>, QK5_1>; return (void*) cpy_q_f32<cpy_blck_q_f32<dequantize_q6_0, QK6_0>, QK6_0>;
} else if (src0->type == GGML_TYPE_F16 && src1->type == GGML_TYPE_F16) { } else if (src0->type == GGML_TYPE_F16 && src1->type == GGML_TYPE_F16) {
return (void*) cpy_flt<cpy_1_flt<half, half>>; return (void*) cpy_flt<cpy_1_flt<half, half>>;
} else if (src0->type == GGML_TYPE_F16 && src1->type == GGML_TYPE_BF16) { } else if (src0->type == GGML_TYPE_F16 && src1->type == GGML_TYPE_BF16) {