Neural net framework built from scratch. Uses gradient descent as learning method. Has graphical support, with buttons to allow easy usage of the feedforward, backpropagate, and gradient_descent functions. Also features nodes that gives instant feedback on the weight,biases, and partial derivatives of the net as well as showing the results of each feed. The quadratic, cross-entropy and log-likelihood cost functions are all included and are customizable. Other features include softmax and the easy export and import of neural nets as a text file. Now supports graphing as well!
AndrewAXue/neural-net
Folders and files
| Name | Name | Last commit date | ||
|---|---|---|---|---|