173 Leaky Integrate Fire Neuron
173 : Leaky Integrate Fire Neuron
- Author: Rocky Lim
- Description: Simulates a Leaky Integrate Fire Neuron based on snnTorch's implementation
- GitHub repository
- Clock: 0 Hz
How it works
This chip takes in an 8-bit number voltage to simulate a Leaky Fire Integrate (LIF) Network. The 8-bit number is split into two different neurons in which they have their respective layers, and it takes that voltage to act as an input current to the LIF neurons. Each neuron generates a spike when the threshold, defined to be 8, is reached or surpassed. Once an input current is passed through, each neuron will decay the value over each clock cycle by shifting the bits of the current state once as it constantly takes the input current. The idea behind the layers is for more significant spikes to be able to reach the output states while less significant events would not affect the output.
How to test
N/A
External hardware
N/A
IO
| # | Input | Output | Bidirectional |
|---|---|---|---|
| 0 | Input Current Bit [0] (Input Neuron 1) | State Variable Bit [0] (Output Neuron 1) | Spike Bit (Output Layer, Neuron 2) |
| 1 | Input Current Bit [1] (Input Neuron 1) | State Variable Bit [1] (Output Neuron 1) | Spike Bit (Output Layer, Neuron 1) |
| 2 | Input Current Bit [2] (Input Neuron 1) | State Variable Bit [2] (Output Neuron 1) | Spike Bit (Inner Layer 2, Neuron 2) |
| 3 | Input Current Bit [3] (Input Neuron 1) | State Variable Bit [3] (Output Neuron 1) | Spike Bit (Inner Layer 2, Neuron 1) |
| 4 | Input Current Bit [4] (Input Neuron 2) | State Variable Bit [4] (Output Neuron 2) | Spike Bit (Inner Layer 1, Neuron 2) |
| 5 | Input Current Bit [5] (Input Neuron 2) | State Variable Bit [5] (Output Neuron 2) | Spike Bit (Inner Layer 1, Neuron 1) |
| 6 | Input Current Bit [6] (Input Neuron 2) | State Variable Bit [6] (Output Neuron 2) | Spike Bit (Input Layer, Neuron 2) |
| 7 | Input Current Bit [7] (Input Neuron 2) | State Variable Bit [7] (Output Neuron 2) | Spike Bit (Input Layer, Neuron 1) |