This is a lightweight AI library that works as a wrapper.
git clone https://github.com/E-Almqvist/ailib.git
- Import it in your project and that is it.
from ailib import ai
my_neural_network = ai.neural_network() # Create an object for your network
my_neural_network.generateLayers( [1, 1] ) # Generate 1 input neuron and 1 output neuron.
my_neural_network.think( [0.2] ) # Make the AI "think" about "0.2" and it will give out 1 output.
# This network does not really do anything usefull.
See example.py
for a better example.
(Checked boxes are implemented features and unchecked has not been implemented yet.)
- Thinking
- Generating neural matrix
- Prediction
- Learning
- Stochastic gradient descent (SGD)
- Relative learning rate
- Support for networks of (neural) networks
- Input/Output passthrough
- Support for loading trained networks
- Cool debug colors
- Multicore processing