Machine Mind

Repository about Machine Learning and Deep Learning.

$ git clone https://github.com/besarthoxhaj/mind.git
$ cd mind/
$ mkdir work
$ docker builder prune
$ docker build . -t mind
$ docker run --rm -p 8888:8888 -v "$PWD"/work:/home/jovyan/work mind
# Entered start.sh with args: jupyter la
# ...
# http://127.0.0.1:8888/lab

Jupyter Notebook

Once inside the Jupyter Notebook select "Python 3" as notebook. Then write and execute a simple command to check everything is ok.

import torch
x = torch.rand(5, 3)
print(x)

Great! Seem everything is working as expected. Let's try now to download the MNIST dataset and run a CNN (Convolutional Neural Network).

from torchvision import datasets
from torchvision.transforms import ToTensor

train_data = datasets.MNIST(
    root = 'data',
    train = True,
    transform = ToTensor(),
    download = True,
)

test_data = datasets.MNIST(
    root = 'data',
    train = False,
    transform = ToTensor()
)

Check if the images are there with matplot library.

import matplotlib.pyplot as plt
plt.imshow(train_data.data[0], cmap='gray')
plt.title('%i' % train_data.targets[0])
plt.show()

If also the last step worked you are all set. Just follow the resources. Or simply search for simple neural networks to do anything, the internet it's full of them.

Resources

1. 3Blue1Brown: https://www.youtube.com/playlist?list=PLZHQObOWTQDNU6R1_67000Dx_ZCJB-3pi
2. PyTorch get started: https://pytorch.org/get-started/locally
3. Google Colab: https://colab.research.google.com
4. PyTorch tutorial: https://pytorch.org/tutorials/beginner/basics/quickstart_tutorial.html