Add Q2_0 quantization: type definition and CPU backend (#24448)

2026-07-07 12:05:47 -07:00
parent c198af4dc2
commit bec4772f6a
17 changed files with 262 additions and 3 deletions
@@ -429,7 +429,8 @@ extern "C" {
        GGML_TYPE_MXFP4   = 39, // MXFP4 (1 block)
        GGML_TYPE_NVFP4   = 40, // NVFP4 (4 blocks, E4M3 scale)
        GGML_TYPE_Q1_0    = 41,
-        GGML_TYPE_COUNT   = 42,
+        GGML_TYPE_Q2_0    = 42,
+        GGML_TYPE_COUNT   = 43,
    };

    // precision
@@ -473,6 +474,7 @@ extern "C" {
        GGML_FTYPE_MOSTLY_MXFP4   = 25, // except 1d tensors
        GGML_FTYPE_MOSTLY_NVFP4   = 26, // except 1d tensors
        GGML_FTYPE_MOSTLY_Q1_0    = 27, // except 1d tensors
+        GGML_FTYPE_MOSTLY_Q2_0    = 28, // except 1d tensors
    };

    // available tensor operations: