def __add__(self, other):
QuantTensor.check_input_type(other)
self.check_scaling_factors_same(other)
output_tensor = self.tensor + other.tensor
output_scale = (self.scale + other.scale) / 2
max_uint_val = max_uint(narrow_range=False, bit_width=self.bit_width)
max_uint_val += max_uint(narrow_range=False, bit_width=other.bit_width)
output_bit_width = ceil_ste(torch.log2(max_uint_val))
output = pack_quant_tensor(output_tensor, output_scale,
output_bit_width)
return output
def __add__(self, other):
QuantTensor.check_input_type(other)
if self.is_valid and other.is_valid:
self.check_scaling_factors_same(other)
output_value = self.value + other.value
output_scale = (self.scale + other.scale) / 2
max_uint_val = max_uint(narrow_range=False, bit_width=self.bit_width)
max_uint_val += max_uint(narrow_range=False, bit_width=other.bit_width)
output_bit_width = ceil_ste(torch.log2(max_uint_val))
output_signed = self.signed or other.signed
output = QuantTensor(output_value, output_scale, output_bit_width, output_signed)
else:
output_value = self.value + other.value
output = QuantTensor(output_value)
return output
def max_output_bit_width(self, input_bit_width, weight_bit_width):
max_input_val = max_uint(bit_width=input_bit_width, narrow_range=False)
max_fc_val = self.weight_quant.tensor_quant.int_quant.max_uint(
weight_bit_width)
max_output_val = max_input_val * max_fc_val * self.in_features
output_bit_width = ceil_ste(torch.log2(max_output_val))
return output_bit_width
def max_acc_bit_width(self, input_bit_width, weight_bit_width):
max_uint_input = max_uint(bit_width=input_bit_width, narrow_range=False)
max_kernel_val = self.weight_quant.max_uint_value(weight_bit_width)
group_size = self.out_channels // self.groups
max_uint_output = max_uint_input * max_kernel_val * self.kernel_size[0] * group_size
max_output_bit_width = ceil_ste(torch.log2(max_uint_output))
return max_output_bit_width
def max_acc_bit_width(self, input_bit_width, weight_bit_width):
max_uint_input = max_uint(bit_width=input_bit_width, narrow_range=False)
max_kernel_val = self.weight_quant.max_uint_value(weight_bit_width)
group_size = self.out_channels // self.groups
overlapping_sums = max(round(self.kernel_size[0] / self.stride[0]), 1)
max_uint_output = max_uint_input * max_kernel_val * overlapping_sums * group_size
max_output_bit_width = ceil_ste(torch.log2(max_uint_output))
return max_output_bit_width
def max_uint(self, bit_width):
return max_uint(self.narrow_range, bit_width)
def max_output_bit_width(self, input_bit_width):
max_input_val = max_uint(bit_width=input_bit_width, narrow_range=False)
max_output_val = max_input_val * self.in_channels
output_bit_width = ceil_ste(torch.log2(max_output_val))
return output_bit_width
def max_output_bit_width(self, input_bit_width):
max_uint_input = max_uint(bit_width=input_bit_width,
narrow_range=False)
max_uint_output = max_uint_input * self.kernel_size * self.kernel_size
max_output_bit_width = ceil_ste(torch.log2(max_uint_output))
return max_output_bit_width
def max_acc_bit_width(self, input_bit_width, reduce_size):
max_uint_input = max_uint(bit_width=input_bit_width,
narrow_range=False)
max_uint_output = max_uint_input * reduce_size
max_output_bit_width = ceil_ste(torch.log2(max_uint_output))
return max_output_bit_width
def max_acc_bit_width(self, input_bit_width):
max_uint_input = max_uint(bit_width=input_bit_width,
narrow_range=False)
max_uint_output = max_uint_input * self._avg_scaling
max_output_bit_width = ceil_ste(torch.log2(max_uint_output))
return max_output_bit_width
def max_acc_bit_width(self, input_bit_width, weight_bit_width):
max_input_val = max_uint(bit_width=input_bit_width, narrow_range=False)
max_weight_val = self.weight_quant.max_uint_value(weight_bit_width)
max_output_val = max_input_val * max_weight_val
output_bit_width = ceil_ste(torch.log2(max_output_val))
return output_bit_width