update circle schema, rename ggml quant data type

Samsung · Aug 28, 2024 · 8634ae4 · 8634ae4
1 parent 3f4f1be
commit 8634ae4
Show file tree

Hide file tree

Showing 6 changed files with 51 additions and 226 deletions.
diff --git a/nnpackage/schema/circle_schema.fbs b/nnpackage/schema/circle_schema.fbs
@@ -68,6 +68,11 @@ enum TensorType : byte {
   UINT32 = 15,
   UINT16 = 16,
   INT4 = 17,
+  // Q4_0, Q4_1, Q8_0, Q8_1 are follow ggml quantization spec (https://github.com/ggerganov/ggml)
+  Q4_0 = 18,
+  Q4_1 = 19,
+  Q8_0 = 20,
+  Q8_1 = 21,
 }
 
 // Custom quantization parameters for experimenting with new quantization
@@ -76,15 +81,9 @@ table CustomQuantization {
   custom:[ubyte] (force_align: 16);
 }
 
-// Block quantization: from ggml quantization (https://github.com/ggerganov/ggml)
-table CircleBlockQuantization {
-  name:string;
-}
-
 // Represents a specific quantization technique's parameters.
 union QuantizationDetails {
-  CustomQuantization,
-  CircleBlockQuantization
+  CustomQuantization
 }
 
 // Parameters for converting a quantized tensor back to float.