Upload folder using huggingface_hub

README.md

@@ -0,0 +1,49 @@
+---
+license: mit
+tags:
+- pytorch
+- safetensors
+- threshold-logic
+- neuromorphic
+- arithmetic
+- multiplier
+---
+
+# threshold-array-multiplier
+
+2x2 array multiplier. Regular structure with rows of AND gates and adders.
+
+## Circuit
+
+```
+Inputs: A[1:0], B[1:0] (4 inputs)
+Outputs: P[3:0] (4 outputs)
+
+      A1    A0
+       │     │
+  B0 ──┼──●──┼──●──► P0
+       │  │  │  │
+  B1 ──●──┼──●──┤
+       │  │     │
+       ▼  ▼     ▼
+      P3 P2    P1
+```
+
+## Array Structure
+
+Regular grid of AND-adder cells. Simple, predictable timing.
+
+## Parameters
+
+| | |
+|---|---|
+| Inputs | 4 |
+| Outputs | 4 |
+| Neurons | 10 |
+| Layers | 3 |
+| Parameters | 44 |
+| Magnitude | 44 |
+
+## License
+
+MIT

config.json

@@ -0,0 +1,9 @@
+{
+  "name": "threshold-array-multiplier",
+  "description": "2x2 array multiplier",
+  "inputs": 4,
+  "outputs": 4,
+  "neurons": 10,
+  "layers": 3,
+  "parameters": 44
+}

create_safetensors.py

@@ -0,0 +1,84 @@
+import torch
+from safetensors.torch import save_file
+
+weights = {}
+
+# 2x2 Array Multiplier
+# Uses regular array structure of AND gates and adders
+# Inputs: A1,A0, B1,B0 (4 inputs)
+# Outputs: P3,P2,P1,P0 (4 outputs)
+
+def add_and(name, idx_a, idx_b, n_inputs):
+    w = [0.0] * n_inputs
+    w[idx_a] = 1.0
+    w[idx_b] = 1.0
+    weights[f'{name}.weight'] = torch.tensor([w], dtype=torch.float32)
+    weights[f'{name}.bias'] = torch.tensor([-2.0], dtype=torch.float32)
+
+def add_xor_direct(name):
+    weights[f'{name}.or.weight'] = torch.tensor([[1.0, 1.0]], dtype=torch.float32)
+    weights[f'{name}.or.bias'] = torch.tensor([-1.0], dtype=torch.float32)
+    weights[f'{name}.nand.weight'] = torch.tensor([[-1.0, -1.0]], dtype=torch.float32)
+    weights[f'{name}.nand.bias'] = torch.tensor([1.0], dtype=torch.float32)
+    weights[f'{name}.and.weight'] = torch.tensor([[1.0, 1.0]], dtype=torch.float32)
+    weights[f'{name}.and.bias'] = torch.tensor([-2.0], dtype=torch.float32)
+
+def add_ha_carry(name):
+    weights[f'{name}.weight'] = torch.tensor([[1.0, 1.0]], dtype=torch.float32)
+    weights[f'{name}.bias'] = torch.tensor([-2.0], dtype=torch.float32)
+
+# Row 0: A[1:0] * B0
+add_and('r0c0', 1, 3, 4)  # A0*B0 -> P0
+add_and('r0c1', 0, 3, 4)  # A1*B0
+
+# Row 1: A[1:0] * B1, shifted and added
+add_and('r1c0', 1, 2, 4)  # A0*B1
+add_and('r1c1', 0, 2, 4)  # A1*B1
+
+# Array cell (0,1): r0c1 + r1c0 -> half adder
+add_xor_direct('cell01_sum')
+add_ha_carry('cell01_carry')
+
+# Array cell (1,1): r1c1 + carry -> half adder
+add_xor_direct('cell11_sum')
+add_ha_carry('cell11_carry')
+
+save_file(weights, 'model.safetensors')
+
+def array_mult(a1, a0, b1, b0):
+    # Partial products
+    r0c0 = a0 & b0  # P0
+    r0c1 = a1 & b0
+    r1c0 = a0 & b1
+    r1c1 = a1 & b1
+
+    # Array addition
+    p0 = r0c0
+    p1 = r0c1 ^ r1c0
+    c1 = r0c1 & r1c0
+    p2 = r1c1 ^ c1
+    p3 = r1c1 & c1
+
+    return p3, p2, p1, p0
+
+print("Verifying 2x2 array multiplier...")
+errors = 0
+for a in range(4):
+    for b in range(4):
+        a1, a0 = (a >> 1) & 1, a & 1
+        b1, b0 = (b >> 1) & 1, b & 1
+        p3, p2, p1, p0 = array_mult(a1, a0, b1, b0)
+        result = p3*8 + p2*4 + p1*2 + p0
+        expected = a * b
+        if result != expected:
+            errors += 1
+            print(f"ERROR: {a}*{b} = {result}, expected {expected}")
+
+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())}")

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ae8072be384f0599d8f6fca68b4a80d079344db2a700d71b4c33886ffe6f557b
+size 1912