fix some errors

src/brevitas/quant/base.py

@@ -357,12 +357,19 @@ class PerChannelPreNorm(ExtendedInjector):
 
 class AccumulatorAwarePerChannelPreNorm(PerChannelPreNorm):
 
-    @value
-    def accumulator_bit_width_impl(accumulator_bit_width):
-        return BitWidthStatefulConst(accumulator_bit_width)
-
     pre_scaling_impl = AccumulatorAwareParameterPreScaling
-    accumulator_bit_width = 32  # default maximum accumulator width is 32 bits
+    accumulator_bit_width = (this << 1).accumulator_bit_width
+    accumulator_bit_width_impl = (this << 1).accumulator_bit_width_impl
+
+
+class AccumulatorAwareZeroCenterPerChannelPreNorm(AccumulatorAwarePerChannelPreNorm):
+
+    pre_scaling_impl = AccumulatorAwareZeroCenterParameterPreScaling
+    pre_zero_point_impl = PreZeroCenterZeroPoint
+    pre_zero_point_shape = this.scaling_shape  # TODO: decouple zero_point from scaling
+    pre_zero_point_stats_input_view_shape_impl = this.scaling_stats_input_view_shape_impl
+    stats_reduce_dim = (this << 1).stats_reduce_dim
+    scaling_shape = (this << 1).scaling_shape
 
 
 class SolvePostScaleGranularity(ExtendedInjector):
@@ -457,10 +464,12 @@ class AccumulatorAwareWeightQuant(WeightNormPerChannelFloatDecoupled):
     per_channel_pre_norm = AccumulatorAwarePerChannelPreNorm
     normalize_stats_impl = PerChannelL1Norm.normalize_stats_impl  # required to align with derivations in paper
     float_to_int_impl = RoundToZeroSte  # required to ensure no upwards rounding violates constraints
+    accumulator_bit_width = 32  # default maximum accumulator width is 32 bits
 
     @value
-    def accumulator_bit_width():
-        return this.per_channel_pre_norm.accumulator_bit_width
+    def accumulator_bit_width_impl(accumulator_bit_width):
+        return BitWidthStatefulConst(accumulator_bit_width)
+
 
 class AccumulatorAwareZeroCenterWeightQuant(AccumulatorAwareWeightQuant):
     """Experimental zero-centered accumulator-aware weight quantized based on:
@@ -470,10 +479,7 @@ class AccumulatorAwareZeroCenterWeightQuant(AccumulatorAwareWeightQuant):
     (1) added zero-centering constraint on the weights (i.e., `PreZeroCenterZeroPoint`)
     (2) a more relaxed l1-norm bound that is derived in the referenced paper
     """
-    pre_scaling_impl = AccumulatorAwareZeroCenterParameterPreScaling
-    pre_zero_point_impl = PreZeroCenterZeroPoint
-    pre_zero_point_shape = this.scaling_shape  # TODO: decouple zero_point from scaling
-    pre_zero_point_stats_input_view_shape_impl = this.scaling_stats_input_view_shape_impl
+    per_channel_pre_norm = AccumulatorAwareZeroCenterPerChannelPreNorm
 
 
 class MSESubInjectorBase(ExtendedInjector):