diff options
| author | Kawrakow <iwankawrakow@gmail.com> | 2024-12-11 11:19:00 +0100 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2024-12-11 11:19:00 +0100 |
| commit | e0adb8b1227dd4622a91a5b3680b4af2e36d32f4 (patch) | |
| tree | 38a28c4e8b948d32495a82dddad6e63d894568a6 /ggml/include/ggml.h | |
| parent | a63a96b5aea031f9dabf1e7c8d5ee28af170e9e7 (diff) | |
Q3_K_R4 (#134)
* q3_k_r4: Zen4 works, but not as good as it should be
238 t/s, so sloghtly slower than q6_k_r4.
* q3_k_r4: NEON
We get PP-512(LLaMA-3.1-8B) = 106.9 t/s.
This is 1.93X faster than q3_K_S!
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
Diffstat (limited to 'ggml/include/ggml.h')
| -rw-r--r-- | ggml/include/ggml.h | 2 |
1 files changed, 2 insertions, 0 deletions
diff --git a/ggml/include/ggml.h b/ggml/include/ggml.h index 7f766497..0ab34f27 100644 --- a/ggml/include/ggml.h +++ b/ggml/include/ggml.h @@ -412,6 +412,7 @@ extern "C" { GGML_TYPE_Q4_0_R4 = 202, GGML_TYPE_Q5_0_R4 = 206, GGML_TYPE_Q8_0_R4 = 208, + GGML_TYPE_Q3_K_R4 = 211, GGML_TYPE_Q4_K_R4 = 212, GGML_TYPE_Q5_K_R4 = 213, GGML_TYPE_Q6_K_R4 = 214, @@ -481,6 +482,7 @@ extern "C" { GGML_FTYPE_MOSTLY_Q4_0_R4 = 202, // except 1d tensors GGML_FTYPE_MOSTLY_Q8_0_R4 = 207, // except 1d tensors GGML_FTYPE_MOSTLY_Q5_0_R4 = 208, // except 1d tensors + GGML_FTYPE_MOSTLY_Q3_K_R4 = 211, // except 1d tensors GGML_FTYPE_MOSTLY_Q4_K_R4 = 212, // except 1d tensors GGML_FTYPE_MOSTLY_Q5_K_R4 = 215, // except 1d tensors GGML_FTYPE_MOSTLY_Q6_K_R4 = 214, // except 1d tensors |