Build Neural network from scratch in go lang for learning purposes

Description

I've created this neural network by randomly generating weights and keeping the baises constant in each layer.The backpropogation works in way where i first calculate the feed forward propoagtion which basically calculates the slope in each layer where i reach to the output layer and then i determine the error by subtracting it with the slope in each layer, then i adjust the weights and baises according to it. For the process, I've refer to this daniel whitenack github to build it https://github.com/dwhitena/gophernet instead of using sigmoid which daniel used i used relu which basically just cut the negative values and reduce the training time, this activation function is widely used as it improve the accuracy significantly

How to run this

first install the iris dataset then

go mod tidy
go run main.go
or
go run .

you can also use build to run the executable just like i did

go build main.go
./main.exe