diff options
| author | Kawrakow <iwankawrakow@gmail.com> | 2024-10-13 13:34:30 +0300 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2024-10-13 13:34:30 +0300 |
| commit | 910a13409463f7aedb0a92be013a1b9bb50f4859 (patch) | |
| tree | 16e13e1fd3010549877408a0a62706b2bc5d5f0c /ggml/src/ggml-cuda/common.cuh | |
| parent | c15de3654e0002537c8052fd6d52d879e778e88c (diff) | |
IQ2_KS: 2.1875 bpw non-linear quantization (#85)
* Experimenting
* iq2k: Try make_qx_quants for the scale
Slightly better for LLaMA-3.1, Gemma-2, slightly worse for
Qwen2.5
* iq2k with make_qx_quants: adjust scale
* iq2ks: basics
* iq2_ks: CUDA works
* iq2_ks: WIP
* iq2_ks: WIP
* iq2_ks: Zen4
* iq2_ks: AVX2
* iq2_ks: scalar dot product
* iq2_ks: ARM_NEON
* iq2_ks: Metal
* iq2_ks: faster Metal
LLaMA-3.1-8B:
PP-512 = 475.22 ± 0.37 t/s
TG-128 = 45.32 ± 0.03 t/s
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
Diffstat (limited to 'ggml/src/ggml-cuda/common.cuh')
| -rw-r--r-- | ggml/src/ggml-cuda/common.cuh | 7 |
1 files changed, 7 insertions, 0 deletions
diff --git a/ggml/src/ggml-cuda/common.cuh b/ggml/src/ggml-cuda/common.cuh index c00cef29..a6a9c3d3 100644 --- a/ggml/src/ggml-cuda/common.cuh +++ b/ggml/src/ggml-cuda/common.cuh @@ -516,6 +516,13 @@ struct ggml_cuda_type_traits<GGML_TYPE_IQ2_K> { }; template<> +struct ggml_cuda_type_traits<GGML_TYPE_IQ2_KS> { + static constexpr int qk = QK_K; + static constexpr int qr = QR4_XS; + static constexpr int qi = QI4_XS; +}; + +template<> struct ggml_cuda_type_traits<GGML_TYPE_IQ3_K> { static constexpr int qk = QK_K; static constexpr int qr = QR4_XS; |