root/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-12-26 03:14:35 +00:00
Code Issues Actions 7 Packages Projects Releases Wiki Activity

Fix type error in quantize_row_q4_1 for Arm NEON

Browse Source
This commit is contained in:
Håkon H. Hitland 2023-04-05 22:59:54 +02:00 committed by Georgi Gerganov
parent 4282f9b0f3
commit bd166f7ffc
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
ggml.c
View File

@ -800,7 +800,7 @@ static void quantize_row_q4_0(const float * restrict x, void * restrict vy, int
const float32x4_t v = vmulq_n_f32(srcv[l], id);
const float32x4_t vf = vaddq_f32(v, vdupq_n_f32(8.5f));
const int32x4_t vi = vcvtq_s32_f32(vf);
const int32x4 vc = vminq_u32(vi, vdupq_n_u32(15));
const int32x4_t vc = vminq_s32(vi, vdupq_n_s32(15));
y[i].qs[2*l + 0] = vgetq_lane_s32(vc, 0) | (vgetq_lane_s32(vc, 1) << 4);
y[i].qs[2*l + 1] = vgetq_lane_s32(vc, 2) | (vgetq_lane_s32(vc, 3) << 4);