NeuralNetworks

NeuralNetworks copied to clipboard

[REMINDER] Notes + some tiny typos

10

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"*

GiorgosXou

I just found out that I can speed up things by using Fixed-point arithmetic, thanks to [this video](https://youtu.be/rY413t5fArw?t=199) at 3:19 *(see also results at 5:00)* * I've also found about...

GiorgosXou

Multiple bias per layer

3

Hi, I'm trying to use your library for TinyML in the Arduino Uno. I have a pre-trained autoencoder model with multiple biases per layer. Do you have an example on...

expeon07

Why not work with 10 bit input data?

4

I have 10 bit input data like this const double inputs[110][8] = { {540,131,48,3,0,0,0,0}, {624,167,63,15,0,0,0,0}, {736,224,96,31,0,0,0,0},... but after learning output is the same for exemple 0.8215888 0.8215888 0.8215888 ... after...

nazariiixa

Save from PROGMEM to EEPROM

1

I just reasiled that for really low-end\cheap MCUs with way too small RAM, it might be a great way to actually save the NN easily in EEPROM, even thought practically...

GiorgosXou

**Depending on the board or MCU** you use, the IDE will either **freeze or** take way too long to **compile**. *(At least with [this build of the old-Arduino](https://archlinux.org/packages/extra/x86_64/arduino/))*. I'm suspecting...

GiorgosXou

NOTE: `NN.print()`'s floating point precision, is inucurate at the last decimal digit.

5

This means that if you copy-paste the printed-weights from the serial output, into another neural-network, there's going to be some slight noise to the output _(when you run it)_. This...

GiorgosXou

Add support for convolutional\kernel\filter layers

2

***My thought proccess:*** I could efficiently add support for convolutional\kernel\filter layers while preserving overall performance via just a fake-activation function and a simple condition when loading weights *(i.e. "activation function"...

GiorgosXou

Resolve warnings: comparison between signed and unsigned integer

2

How I messed-up signed and unsigned... ``` home/xou/Arduino/libraries/NeuralNetworks/src/NeuralNetwork.h: In member function 'void NeuralNetwork::pdestract()': /home/xou/Arduino/libraries/NeuralNetworks/src/NeuralNetwork.h:1096:35: warning: comparison between signed and unsigned integer expressions [-Wsign-compare] for (int i = 0; i <...

GiorgosXou

I just forgot to append CA2... CA5: https://github.com/GiorgosXou/NeuralNetworks/blob/68670d51d5bd1da26b6ae14707ebec4abb0c276a/src/NeuralNetwork.h#L694

GiorgosXou