Use with credit: (Neural Network, by @0MXC) 5 hours to code SPECIFICATIONS ⌜‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾⌝ 4 Lists total (Layers, Node IDX, Weights, Biases) 1 list for all layers (Input, Hidden, Output) 126 blocks (with visuals) and 85 blocks (without visuals) 0.003 sec to run with these settings * (shown in diagram) Uses 1/(1+e^-x) sigmoid activation function Fully customizable weight and bias value range Fully customizable amount of neurons in each layer Fully customizable amount of hidden layers 20+ Digits of accuracy *2 Inputs, 2 Hiden layers (5 neurons each), 2 Outputs ⌞___________________________________________⌟ EDITING THE ENGINE Editing layer structure: You can use the Node IDX list to edit layer info. The first list entry is the number of inputs, then every entry following except the last will be counted as a hidden layer, the last is counted as the output layer. Editing the neurons per layer: Edit the value of each of the entries. Editing bias and weight value ranges: See the comments in the project. Editing inputs: Follow the comments then duplicate and add your values. Additionally, I recommend deleting the code designated as visual code as that is just for the display and is not needed. DISCLAIMER I understand how Neural Networks function, but I do not know what values are best so in no way are the values set in the project perfect or even correct. Leave me a comment with the values that you used to make it, with an explanation of why it works and I might update them. Finally, I am not perfect and may make mistakes in the engine, just let me know about them and I will fix them. UPDATE LOG v0.0-Shared v0.1-
