Hi, I built a convolutional neural network in PyTorch and trained it to classify characters. I created it to have 4 convolutional layers, 2 max pooling layers, and 1 linear layer, and it was inspired by the pre-existing TinyVGG CNN architecture. It took about 40 minutes to train with the EMNIST dataset on Google Colab.
You can try it yourself with the live demo on my website here.
Check out the Google Colab notebook I built and trained it on here! (or look at the .ipynb file in this repo)
Thanks for checking it out, take care 👋
