gguf-py: Add IQ1_M to GGML_QUANT_SIZES

pmysl · pmysl · commit 5622e3aa2263 · 2024-04-19T10:40:54.000+02:00
diff --git a/gguf-py/gguf/constants.py b/gguf-py/gguf/constants.py
@@ -859,6 +859,7 @@ def get_type(val: Any) -> GGUFValueType:
     GGMLQuantizationType.I32:     (1, 4),
     GGMLQuantizationType.I64:     (1, 8),
     GGMLQuantizationType.F64:     (1, 8),
+    GGMLQuantizationType.IQ1_M:   (256, QK_K // 8 + QK_K // 16  + QK_K // 32),
 }
 
 

Original file line number	Diff line number	Diff line change
`@@ -859,6 +859,7 @@ def get_type(val: Any) -> GGUFValueType:`
`859`	`859`	`GGMLQuantizationType.I32: (1, 4),`
`860`	`860`	`GGMLQuantizationType.I64: (1, 8),`
`861`	`861`	`GGMLQuantizationType.F64: (1, 8),`
	`862`	`+ GGMLQuantizationType.IQ1_M: (256, QK_K // 8 + QK_K // 16 + QK_K // 32),`
`862`	`863`	`}`
`863`	`864`
`864`	`865`