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 /include/llama.h | |
| parent | 291066e6df5318c322a03e592483aae8820d3b19 (diff) | |
iq2_k: Basics
Quantize/dequantize, CUDA deqantize, AVX512 iqk_mul_mat.
Diffstat (limited to 'include/llama.h')
| -rw-r--r-- | include/llama.h | 5 |
1 files changed, 3 insertions, 2 deletions
diff --git a/include/llama.h b/include/llama.h index a90b56e1..3549d3f3 100644 --- a/include/llama.h +++ b/include/llama.h @@ -168,9 +168,10 @@ extern "C" { LLAMA_FTYPE_MOSTLY_Q4_0_4_4 = 33, // except 1d tensors LLAMA_FTYPE_MOSTLY_Q4_0_4_8 = 34, // except 1d tensors LLAMA_FTYPE_MOSTLY_Q4_0_8_8 = 35, // except 1d tensors - LLAMA_FTYPE_MOSTLY_IQ1_BN = 36, - LLAMA_FTYPE_MOSTLY_IQ2_BN = 37, + LLAMA_FTYPE_MOSTLY_IQ1_BN = 36, // except 1d tensors + LLAMA_FTYPE_MOSTLY_IQ2_BN = 37, // except 1d tensors LLAMA_FTYPE_MOSTLY_IQ4_K = 38, // except 1d tensors + LLAMA_FTYPE_MOSTLY_IQ2_K = 39, // except 1d tensors LLAMA_FTYPE_GUESSED = 1024, // not specified in the model file }; |