ATtiny85 mynn POC

Proof of concept for using my hobby built mynn neural network library on an ATtiny85.

Simple pre-defined model approximating the output of a XOR gate, feeding it data read from input pins and assigning the output pins based on the output of the network.

Build Instructions

1. Install prerequisites as described in the avr-hal readme (avr-gcc, avr-libc, avrdude).

2. Run cargo build --release to build the firmware.

3. Use an Arduino as an ISP to upload your firmware:

   • Connect your AVR based Arduino to your PC, opening the Arduino IDE, select the correct port and Arduino board then open the example sketch File -> Examples -> 11.ArduinoISP -> ArduinoISP and upload it.

   • Take note of what port your arduino is on

4. Create your circuit:

   • Arduino ISP circuit (used for programming):
     • Arduino pin 13 -> ATtiny85 pin 7
     • Arduino pin 12 -> ATtiny85 pin 6
     • Arduino pin 11 -> ATtiny85 pin 5
     • Arduino pin 10 -> ATtiny85 pin 1
   • Arduino 5V -> ATtiny pin 8
   • Arduino GND -> ATtiny pin 4
   • LED between pin 3 and pin 4 (GND) on the ATtiny85 (optional resistor)
   • Jumper wires on pins 5 & 6 on the ATtiny85 (these are you 2 XOR inputs)

5. With avrdude installed, upload using the following command:

avrdude -c arduino_as_isp -p t85 -P [serial port of arduino] -U flash:w:"target/avr-attiny85/release/[executable name].elf":a

Operation

When you connect either pin 5 or 6 to ground, the LED should turn on, when they are both connected or disconected to ground the LED will turn off, this is because the program is running a constant loop reading the 2 inputs through a neural network trained to approximate a XOR gate (turns on when both of it's inputs are different) and then assigning the output pin based on it's approximated result.