root/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-11-11 13:30:35 +00:00
Code Issues Actions 11 Packages Projects Releases Wiki Activity

Avoid division-by-zero on 0-weights

Browse Source
This commit is contained in:
Sigbjørn Skjæret 2024-06-07 23:57:04 +02:00 committed by GitHub
parent da799b4189
commit 0225cfbfb3
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
ggml-quants.c
View File

@ -1754,7 +1754,7 @@ static float make_q3_quants(int n, int nmax, const float * restrict x, int8_t *
for (int i = 0; i < n; ++i) {
L[i] += nmax;
}
return sumlx / suml2;
return suml2 ? sumlx / suml2 : 0.0f;
}
for (int i = 0; i < n; ++i) {
int l = nearest_int(iscale * x[i]);
@ -2158,7 +2158,7 @@ static float make_qp_quants(int n, int nmax, const float * restrict x, uint8_t *
break;
}
}
return sumlx/suml2;
return suml2 ? sumlx/suml2 : 0.0f;
}
static void quantize_row_q2_K_impl(const float * restrict x, block_q2_K * restrict y, int k, const float * restrict quant_weights) {