File size: 2,955 Bytes
698e230 | 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 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 | import torch
from safetensors.torch import save_file
weights = {}
# BCH(7,4) Decoder / Hamming(7,4) Decoder
# Computes syndrome and passes through data bits
def add_neuron(name, w_list, bias):
weights[f'{name}.weight'] = torch.tensor([w_list], dtype=torch.float32)
weights[f'{name}.bias'] = torch.tensor([bias], dtype=torch.float32)
# Input: R6, R5, R4, R3, R2, R1, R0 (7-bit received word)
# Output: D3, D2, D1, D0 (4 data bits), S2, S1, S0 (syndrome)
# Pass through data bits
add_neuron('d3', [1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], -1.0) # R6
add_neuron('d2', [0.0, 1.0, 0.0, 0.0, 0.0, 0.0, 0.0], -1.0) # R5
add_neuron('d1', [0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0], -1.0) # R4
add_neuron('d0', [0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0], -1.0) # R3
# Syndrome bits (parity checks)
# S0 checks positions 1,3,5,7 -> R6,R4,R2,R0 (indices 0,2,4,6)
add_neuron('s0_at1', [1.0, 0.0, 1.0, 0.0, 1.0, 0.0, 1.0], -1.0)
add_neuron('s0_at2', [1.0, 0.0, 1.0, 0.0, 1.0, 0.0, 1.0], -2.0)
add_neuron('s0_at3', [1.0, 0.0, 1.0, 0.0, 1.0, 0.0, 1.0], -3.0)
add_neuron('s0_at4', [1.0, 0.0, 1.0, 0.0, 1.0, 0.0, 1.0], -4.0)
# S1 checks positions 2,3,6,7 -> R6,R5,R2,R1 (indices 0,1,4,5)
add_neuron('s1_at1', [1.0, 1.0, 0.0, 0.0, 1.0, 1.0, 0.0], -1.0)
add_neuron('s1_at2', [1.0, 1.0, 0.0, 0.0, 1.0, 1.0, 0.0], -2.0)
add_neuron('s1_at3', [1.0, 1.0, 0.0, 0.0, 1.0, 1.0, 0.0], -3.0)
add_neuron('s1_at4', [1.0, 1.0, 0.0, 0.0, 1.0, 1.0, 0.0], -4.0)
# S2 checks positions 4,5,6,7 -> R6,R5,R4,R2 (indices 0,1,2,4)
add_neuron('s2_at1', [1.0, 1.0, 1.0, 0.0, 1.0, 0.0, 0.0], -1.0)
add_neuron('s2_at2', [1.0, 1.0, 1.0, 0.0, 1.0, 0.0, 0.0], -2.0)
add_neuron('s2_at3', [1.0, 1.0, 1.0, 0.0, 1.0, 0.0, 0.0], -3.0)
add_neuron('s2_at4', [1.0, 1.0, 1.0, 0.0, 1.0, 0.0, 0.0], -4.0)
save_file(weights, 'model.safetensors')
def xor4(a, b, c, d):
return a ^ b ^ c ^ d
def bch_decode(r6, r5, r4, r3, r2, r1, r0):
# Syndrome computation
s0 = xor4(r6, r4, r2, r0)
s1 = xor4(r6, r5, r2, r1)
s2 = xor4(r6, r5, r4, r2)
# Data bits (without correction for simplicity)
d3, d2, d1, d0 = r6, r5, r4, r3
return d3, d2, d1, d0, s2, s1, s0
print("Verifying BCH(7,4) decoder...")
errors = 0
# Test with valid codewords
def encode(d3, d2, d1, d0):
c2 = d3 ^ d2 ^ d0
c1 = d3 ^ d1 ^ d0
c0 = d2 ^ d1 ^ d0
return d3, d2, d1, d0, c2, c1, c0
for d in range(16):
d3, d2, d1, d0 = (d>>3)&1, (d>>2)&1, (d>>1)&1, d&1
codeword = encode(d3, d2, d1, d0)
decoded = bch_decode(*codeword)
# Check data extraction
if decoded[:4] != (d3, d2, d1, d0):
errors += 1
print(f"Data error for d={d}")
if errors == 0:
print("All 16 test cases passed!")
else:
print(f"FAILED: {errors} errors")
mag = sum(t.abs().sum().item() for t in weights.values())
print(f"Magnitude: {mag:.0f}")
print(f"Parameters: {sum(t.numel() for t in weights.values())}")
|