mirror of
https://github.com/ggml-org/llama.cpp.git
synced 2025-08-14 20:29:41 -04:00
ggml : fix cpy op for IQ-quants to use reference impl
ggml-ci
This commit is contained in:
@@ -2968,8 +2968,12 @@ static void ggml_compute_forward_dup_f16(
|
||||
id += ne00 * (ne01 - ir1);
|
||||
}
|
||||
}
|
||||
} else if (ggml_get_type_traits_cpu(dst->type)->from_float) {
|
||||
ggml_from_float_t const quantize_row_q = ggml_get_type_traits_cpu(dst->type)->from_float;
|
||||
} else if (ggml_get_type_traits_cpu(dst->type)->from_float || ggml_get_type_traits(dst->type)->from_float_ref) {
|
||||
ggml_from_float_t quantize_row_q = ggml_get_type_traits_cpu(dst->type)->from_float;
|
||||
if (!quantize_row_q) {
|
||||
quantize_row_q = ggml_get_type_traits(dst->type)->from_float_ref;
|
||||
}
|
||||
|
||||
float * src0_f32 = (float *) params->wdata + (ne00 + CACHE_LINE_SIZE_F32) * ith;
|
||||
|
||||
size_t id = 0;
|
||||
@@ -3565,8 +3569,11 @@ static void ggml_compute_forward_dup_f32(
|
||||
id += rs * (ne01 - ir1);
|
||||
}
|
||||
}
|
||||
} else if (ggml_get_type_traits_cpu(dst->type)->from_float) {
|
||||
ggml_from_float_t const quantize_row_q = ggml_get_type_traits_cpu(dst->type)->from_float;
|
||||
} else if (ggml_get_type_traits_cpu(dst->type)->from_float || ggml_get_type_traits(dst->type)->from_float_ref) {
|
||||
ggml_from_float_t quantize_row_q = ggml_get_type_traits_cpu(dst->type)->from_float;
|
||||
if (!quantize_row_q) {
|
||||
quantize_row_q = ggml_get_type_traits(dst->type)->from_float_ref;
|
||||
}
|
||||
|
||||
size_t id = 0;
|
||||
size_t rs = nb0 * (ne00 / ggml_blck_size(dst->type));
|
||||
|
Reference in New Issue
Block a user