diff options
| author | Iwan Kawrakow <iwan.kawrakow@gmail.com> | 2024-07-28 19:43:18 +0300 |
|---|---|---|
| committer | Kawrakow <48489457+ikawrakow@users.noreply.github.com> | 2024-08-01 09:38:06 +0200 |
| commit | c85e139c68b0b0b15656cc0bc9618d632ed18822 (patch) | |
| tree | f07e1f8c59522d286301c45de78d52ba48353ac4 /ggml/include | |
| parent | 291066e6df5318c322a03e592483aae8820d3b19 (diff) | |
iq2_k: Basics
Quantize/dequantize, CUDA deqantize, AVX512 iqk_mul_mat.
Diffstat (limited to 'ggml/include')
| -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 ff7f0064..2cb4af32 100644 --- a/ggml/include/ggml.h +++ b/ggml/include/ggml.h @@ -390,6 +390,7 @@ extern "C" { GGML_TYPE_IQ2_BN = 35, GGML_TYPE_Q8_K64 = 36, GGML_TYPE_IQ4_K = 37, + GGML_TYPE_IQ2_K = 38, GGML_TYPE_COUNT, }; @@ -437,6 +438,7 @@ extern "C" { GGML_FTYPE_MOSTLY_IQ1_BN = 28, // except 1d tensors GGML_FTYPE_MOSTLY_IQ2_BN = 29, // except 1d tensors GGML_FTYPE_MOSTLY_IQ4_K = 30, // except 1d tensors + GGML_FTYPE_MOSTLY_IQ2_K = 31, // except 1d tensors }; // available tensor operations: |