Open GiorgosXou opened 2 years ago
https://github.com/GiorgosXou/NeuralNetworks/blob/31bb95e4cefcaefb28f6b031a634ded6c7274a7d/examples/Backpropagation_double_Xor/Backpropagation_double_Xor.ino#L2
https://github.com/GiorgosXou/NeuralNetworks/blob/dee00f691357a4712ea981b6f48e84e290c4df90/examples/Any_Activation_Function_Per_Layer/Any_Activation_Function_Per_Layer.ino#L3
^ False link, https://github.com/GiorgosXou/NeuralNetworks#define-macro-properties
In cases of any change to the examples, reminder to update:
In the readme.md, I have cutstom instead of custom in a title
In the readme.md, Summary > Output column, change float
to DFLOAT
typo
Nothing too important, just a visualization thing, make those horizontal
This softmax
part-implementation for use with any ACTIVATION__PER_LAYER
, (even thought I trust myself, not gonna lie) seems slightly suspicious:
https://github.com/GiorgosXou/NeuralNetworks/blob/da5b5af6d0f744c5c25b4879ab0987bf94318c0d/src/NeuralNetwork.h#L2041
And by suspicious, I mean: i don't remember 😠+ division lol
Reminder: Remove space in middle after =
:
https://github.com/GiorgosXou/NeuralNetworks/blob/da5b5af6d0f744c5c25b4879ab0987bf94318c0d/src/NeuralNetwork.h#L1509
I forgot to remove it, when I was copying a previous example, lol: https://github.com/GiorgosXou/NeuralNetworks/blob/da5b5af6d0f744c5c25b4879ab0987bf94318c0d/examples/Media/Save_NN_to_internal_EEPROM/Save_NN_to_internal_EEPROM.ino#L14-L24
On Arduino Uno usually you need to specify CS
pin to whatever you have it, usually begin(4)
, it's nice to have as a comment
https://github.com/GiorgosXou/NeuralNetworks/blob/da5b5af6d0f744c5c25b4879ab0987bf94318c0d/examples/Media/Save_load_NN_from_SD/Save_load_NN_from_SD.ino#L41
AAAAAAAAAAAAAAA, sorry but i have an error that fried my head and I can't even write propperly lol
https://github.com/GiorgosXou/NeuralNetworks/blob/dee00f691357a4712ea981b6f48e84e290c4df90/examples/FeedForward_double_Xor/FeedForward_double_Xor.ino#L1
calculates the number of "elements in an array" not "layers"*