model.py

import torch
from safetensors.torch import load_file

def load_model(path='model.safetensors'):
    return load_file(path)

def thermometer_to_binary(therm, weights):
    """Convert 7-bit thermometer to 3-bit binary.
    Returns (b2, b1, b0).
    """
    t = torch.tensor([float(x) for x in therm])

    # b2 = t3 (direct)
    b2 = int((t * weights['b2.weight']).sum() + weights['b2.bias'] >= 0)

    # b1 = (t1 AND NOT(t3)) OR t5
    and_result = int((t * weights['b1_and.weight']).sum() + weights['b1_and.bias'] >= 0)
    b1_inputs = torch.tensor([float(and_result), float(therm[5])])
    b1 = int((b1_inputs * weights['b1_or.weight']).sum() + weights['b1_or.bias'] >= 0)

    # b0 = (t0 AND NOT(t1)) OR (t2 AND NOT(t3)) OR (t4 AND NOT(t5)) OR t6
    and0 = int((t * weights['b0_and0.weight']).sum() + weights['b0_and0.bias'] >= 0)
    and2 = int((t * weights['b0_and2.weight']).sum() + weights['b0_and2.bias'] >= 0)
    and4 = int((t * weights['b0_and4.weight']).sum() + weights['b0_and4.bias'] >= 0)
    b0_inputs = torch.tensor([float(and0), float(and2), float(and4), float(therm[6])])
    b0 = int((b0_inputs * weights['b0_or.weight']).sum() + weights['b0_or.bias'] >= 0)

    return b2, b1, b0

if __name__ == '__main__':
    w = load_model()
    print('Thermometer to Binary Converter')
    thermometers = [
        [0,0,0,0,0,0,0], [1,0,0,0,0,0,0], [1,1,0,0,0,0,0], [1,1,1,0,0,0,0],
        [1,1,1,1,0,0,0], [1,1,1,1,1,0,0], [1,1,1,1,1,1,0], [1,1,1,1,1,1,1],
    ]
    for therm in thermometers:
        b2, b1, b0 = thermometer_to_binary(therm, w)
        print(f"{''.join(map(str,therm))} -> {b2*4 + b1*2 + b0}")