https://github.com/armandpl/nn_from_scratch
Neural Net from scratch
https://github.com/armandpl/nn_from_scratch
neural-networks numpy python
Last synced: 10 months ago
JSON representation
Neural Net from scratch
- Host: GitHub
- URL: https://github.com/armandpl/nn_from_scratch
- Owner: Armandpl
- Created: 2019-11-15T14:11:03.000Z (about 6 years ago)
- Default Branch: master
- Last Pushed: 2020-09-13T22:50:22.000Z (over 5 years ago)
- Last Synced: 2024-12-17T16:48:03.307Z (about 1 year ago)
- Topics: neural-networks, numpy, python
- Language: Python
- Homepage:
- Size: 18.7 MB
- Stars: 0
- Watchers: 2
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Neural Net from scratch
This is my first atempt at implementing a neural net from scratch, without looking at any code.
# Useful ressources:
http://neuralnetworksanddeeplearning.com/chap2.html
[Neural Networks - 3b1b](https://www.youtube.com/watch?v=aircAruvnKk&list=PLZHQObOWTQDNU6R1_67000Dx_ZCJB-3pi)
https://towardsdatascience.com/weight-initialization-techniques-in-neural-networks-26c649eb3b78
## Todo
* finish implementing multiple activation functions
* implement one cycle policy
* implement multiple loss functions
* implement multiple layer types
* implement multiple metrics
* numpy exp overflow ?
* leaky relu ?