python cleanup

Author: horasal

gguf-py/gguf/quants.py

@@ -15,6 +15,7 @@ def e8m0_to_fp32_any(x: np.ndarray, e: np.uint32) -> np.ndarray:
     bits = np.where(x < e + 1, np.uint32(1) << np.uint32(x - e + 22), np.uint32(x - e) << np.uint32(23))
     return bits.view(np.float32)
 
+
 def quant_shape_to_byte_shape(shape: Sequence[int], quant_type: GGMLQuantizationType) -> tuple[int, ...]:
     block_size, type_size = GGML_QUANT_SIZES[quant_type]
     if shape[-1] % block_size != 0:
@@ -730,9 +731,7 @@ def quantize_blocks(cls, blocks: np.ndarray) -> np.ndarray:
 
         with np.errstate(divide="ignore"):
             # convert log2(d_max) to e8m0
-            e = np.where(d_max > 0, np.floor(np.log2(d_max)) - 9 + 127, 0).astype(
-                np.uint8
-            )
+            e = np.where(d_max > 0, np.floor(np.log2(d_max)) - 3 + 127, 0).astype(np.uint8)
 
         # d is float of above e8m0
         d = cls.__e8m0_to_fp32_scaled(e)
@@ -817,9 +816,7 @@ def quantize_blocks(cls, blocks: np.ndarray) -> np.ndarray:
 
         with np.errstate(divide="ignore"):
             # convert log2(d_max) to e8m0
-            e = np.where(d_max > 0, np.floor(np.log2(d_max)) - 6 + 127, 0).astype(
-                np.uint8
-            )
+            e = np.where(d_max > 0, np.floor(np.log2(d_max)) - 1 + 127, 0).astype(np.uint8)
 
         # d is float of above e8m0
         d = cls.__e8m0_to_fp32_scaled(e)