File size: 1,531 Bytes
0443898 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 |
"""
Threshold Network for MOD-4 Circuit
A formally verified threshold network computing Hamming weight mod 4.
Uses the algebraic weight pattern [1, 1, 1, -3, 1, 1, 1, -3].
"""
import torch
from safetensors.torch import load_file
class ThresholdMod4:
"""
MOD-4 circuit using threshold logic.
Weight pattern: (1, 1, 1, 1-m) repeating for m=4
Computes cumulative sum that cycles mod 4.
"""
def __init__(self, weights_dict):
self.weight = weights_dict['weight']
self.bias = weights_dict['bias']
def __call__(self, bits):
inputs = torch.tensor([float(b) for b in bits])
weighted_sum = (inputs * self.weight).sum() + self.bias
return weighted_sum
def get_residue(self, bits):
"""Returns Hamming weight mod 4."""
return sum(bits) % 4
@classmethod
def from_safetensors(cls, path="model.safetensors"):
return cls(load_file(path))
def forward(x, weights):
x = torch.as_tensor(x, dtype=torch.float32)
weighted_sum = (x * weights['weight']).sum(dim=-1) + weights['bias']
return weighted_sum
if __name__ == "__main__":
weights = load_file("model.safetensors")
model = ThresholdMod4(weights)
print("MOD-4 Circuit Tests:")
print("-" * 40)
for hw in range(9):
bits = [1]*hw + [0]*(8-hw)
out = model(bits).item()
expected_residue = hw % 4
print(f"HW={hw}: weighted_sum={out:.0f}, HW mod 4 = {expected_residue}")
|