Binary neural network
Weights and activations in {-1, +1}. A neuron is popcount(xnor(x, w)) against a threshold. Training is seeded hill-climbing on the full table, so the same task, width and seed always give the same net.
deterministic, in this tab, nothing sent
Truth table
Inputs on the left, target on the right. On the custom task, click a target cell to flip it.
Netlist
unverifiedNAND–
LATCH–
Transistors–
Net accuracy–
Vectors verified–
- Netlist hash
- –
- Canonical form
- topological order, dense ids, keccak256 of the JSON below
Canonical netlist
Verilog
Tape out
not connectedBurns the bill of materials from your wallet to the burn address, with the record behind the transfer. One signature. The transaction hash is the Machine ID.
Your transistors
–
connect to read the balance- Read the token and the balance
- Simulate the burn with the record attached
- Sign in the wallet
- Mined