README.md

---
license: mit
tags:
- pytorch
- safetensors
- threshold-logic
- neuromorphic
- sequential
- latch
---

# threshold-sr-latch

SR (Set-Reset) latch next-state logic as threshold circuit. Computes next state given current inputs and previous state.

## Circuit

```
S      ──┐
R      ──┼──► SR Latch ──┬──► Q
Q_prev ──┘               └──► Qn
```

## Truth Table

| S | R | Q_prev | Q | Qn | Mode |
|---|---|--------|---|----|----|
| 0 | 0 | 0 | 0 | 1 | Hold |
| 0 | 0 | 1 | 1 | 0 | Hold |
| 0 | 1 | X | 0 | 1 | Reset |
| 1 | 0 | X | 1 | 0 | Set |
| 1 | 1 | X | 0 | 0 | Invalid |

## Classic NOR Implementation

```
     S ──┬──►[NOR]──► Qn
         │      ▲
         │      │
         │      └─────┐
         │            │
     R ──┴──►[NOR]──► Q
              ▲       │
              │       │
              └───────┘
```

Cross-coupled NOR gates create feedback loop.

## Combinational Representation

This circuit models the **next-state function**:
- Input: S, R, and Q_prev (previous Q)
- Output: Q_next, Qn_next

True latch behavior requires feeding Q output back to Q_prev over time.

## Architecture

| Component | Neurons |
|-----------|---------|
| Input inversions | 6 |
| Hold logic | 2 |
| Output gates | 3 |

**Total: 11 neurons, 39 parameters, 4 layers**

## Usage

```python
from safetensors.torch import load_file

w = load_file('model.safetensors')

# Simulate latch over time:
q = 0
for s, r in [(1,0), (0,0), (0,1), (0,0)]:
    q_next = compute_q(s, r, q, w)
    q = q_next
```

## Files

```
threshold-sr-latch/
├── model.safetensors
├── create_safetensors.py
├── config.json
└── README.md
```

## License

MIT