Skip to content

ismlrn/cnn-scratch

Repository files navigation

CNN from Scratch – NumPy (without using PyTorch or TensorFlow)

a full implementation of a simple Convolutional Neural Network using only NumPy.

File

  • cnn-full.py – Implements:
    • conv2d (padding, arbitrary filters)
    • relu
    • maxpool2d
    • flatten
    • dense (fully connected layer)
    • softmax

How to Run

You can use cnn-full.ipynb or:

pip install numpy
python main.py

About

NumPy-only CNN implementation - Implements conv2d, ReLU, maxpool2d, flatten, dense, softmax and backpropagating from scratch

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors