nikola/llama_cpp
1
0
Fork 0
Code Issues Pull Requests Actions 33 Packages Projects Releases Wiki Activity

ggml: add Q1_0 1-bit quantization support (CPU) (#21273)

Browse Source 
* ggml: add Q1_0 and Q1_0_g128 1-bit quantization support (CPU)

* add generic fallback for x86

* remove Q1_0 (group size 32)

* rename Q1_0_g128 => Q1_0

* fix Q1_0 LlamaFileType Enum

* Fix trailing spaces; add generic fallback for othre backends

* Apply suggestions from code review

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>

* fix /r/n spacing + arch-fallback

---------

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>
This commit is contained in:
Pasha Khosravi
2026-04-06 20:55:21 +02:00
committed by GitHub
co-authored by Sigbjørn Skjæret
parent 506200cf8b
commit 2e1f0a889e
21 changed files with 285 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ggml/src/ggml.c
+10
View File
@@ -651,6 +651,14 @@ static const struct ggml_type_traits type_traits[GGML_TYPE_COUNT] = {
.to_float = (ggml_to_float_t) ggml_fp16_to_fp32_row,
.from_float_ref = (ggml_from_float_t) ggml_fp32_to_fp16_row,
},
[GGML_TYPE_Q1_0] = {
.type_name = "q1_0",
.blck_size = QK1_0,
.type_size = sizeof(block_q1_0),
.is_quantized = true,
.to_float = (ggml_to_float_t) dequantize_row_q1_0,
.from_float_ref = (ggml_from_float_t) quantize_row_q1_0_ref,
},
[GGML_TYPE_Q4_0] = {
.type_name = "q4_0",
.blck_size = QK4_0,
@@ -1384,6 +1392,7 @@ enum ggml_type ggml_ftype_to_ggml_type(enum ggml_ftype ftype) {
case GGML_FTYPE_MOSTLY_BF16: wtype = GGML_TYPE_BF16; break;
case GGML_FTYPE_MOSTLY_Q4_0: wtype = GGML_TYPE_Q4_0; break;
case GGML_FTYPE_MOSTLY_Q4_1: wtype = GGML_TYPE_Q4_1; break;
case GGML_FTYPE_MOSTLY_Q1_0: wtype = GGML_TYPE_Q1_0; break;
case GGML_FTYPE_MOSTLY_Q5_0: wtype = GGML_TYPE_Q5_0; break;
case GGML_FTYPE_MOSTLY_Q5_1: wtype = GGML_TYPE_Q5_1; break;
case GGML_FTYPE_MOSTLY_Q8_0: wtype = GGML_TYPE_Q8_0; break;
@@ -7652,6 +7661,7 @@ size_t ggml_quantize_chunk(
size_t result = 0;
switch (type) {
case GGML_TYPE_Q1_0: result = quantize_q1_0(src + start, (char *) dst + start_row * row_size, nrows, n_per_row, imatrix); break;
case GGML_TYPE_Q4_0: result = quantize_q4_0(src + start, (char *) dst + start_row * row_size, nrows, n_per_row, imatrix); break;
case GGML_TYPE_Q4_1: result = quantize_q4_1(src + start, (char *) dst + start_row * row_size, nrows, n_per_row, imatrix); break;
case GGML_TYPE_Q5_0: result = quantize_q5_0(src + start, (char *) dst + start_row * row_size, nrows, n_per_row, imatrix); break;