ikawrakow/ik_llama.cpp
1
0
Fork 0
mirror of https://github.com/ikawrakow/ik_llama.cpp.git synced 2026-05-01 03:41:53 +00:00
Code Issues Packages Projects Releases Wiki Activity

IQ3_K_R4 (#145)

Browse Source 
* iq3_k_r4 WIP

* iq3_k_r4: Zen4

* iq3_k_r4: AVX2

* iq3_k_r4: NEON

* iq3_k_r4: faster matrix x vector multiplication on NEON

---------

Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
This commit is contained in:
Kawrakow
2024-12-17 07:51:11 +01:00
committed by GitHub
parent 1095cd27a9
commit b52e2e2934
10 changed files with 482 additions and 39 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
ggml/include/ggml.h
View File

@@ -423,6 +423,7 @@ extern "C" {
GGML_TYPE_BF16_R16 = 230,
GGML_TYPE_Q6_0_R4 = 233,
GGML_TYPE_IQ2_BN_R4 = 335,
GGML_TYPE_IQ3_K_R4 = 338,
GGML_TYPE_IQ4_K_R4 = 339,
GGML_TYPE_Q8_K_R8 = 399,
GGML_TYPE_COUNT,
@@ -497,6 +498,7 @@ extern "C" {
GGML_FTYPE_MOSTLY_BF16_R16 = 224, // except 1d tensors
GGML_FTYPE_MOSTLY_Q6_0_R4 = 227, // except 1d tensors
GGML_FTYPE_MOSTLY_IQ2_BN_R4 = 329, // except 1d tensors
GGML_FTYPE_MOSTLY_IQ3_K_R4 = 331, // except 1d tensors
GGML_FTYPE_MOSTLY_IQ4_K_R4 = 332, // except 1d tensors
GGML_FTYPE_MOSTLY_Q8_K_R8 = 399, // except 1d tensors
};