vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Remove Q_K 0 type changes

Browse source
This commit is contained in:
Henrik Forstén 2023-12-28 17:13:39 +02:00
parent 3fafecae9e
commit 2a21bd8ab2
1 changed files with 0 additions and 4 deletions
Download patch file Download diff file Expand all files Collapse all files

4
ggml-quants.c
View file

@ -1754,8 +1754,6 @@ void quantize_row_q3_K_reference(const float * restrict x, block_q3_K * restrict
}
#endif
y[i].d = GGML_FP32_TO_FP16(lstsq_q_0(q_fit, x, QK_K));
memset(y[i].hmask, 0, QK_K/8);
// We put the high-bit for the 1st 8 quants into bit 0, the next 8 into bit 1, etc.
int m = 0;
@ -2232,8 +2230,6 @@ void quantize_row_q6_K_reference(const float * restrict x, block_q6_K * restrict
}
}
y[i].d = GGML_FP32_TO_FP16(lstsq_q_0(q_fit, x, QK_K));
uint8_t * restrict ql = y[i].ql;
uint8_t * restrict qh = y[i].qh;
#if QK_K == 256