In this notebook, we will learn to experiment solving MNIST dataset using several different learning algorithm e.g. Logistic Regression, Random Forest, Multi Layer Perceptron (MLP), and Convolutional Neural Network (CNN) while tracking the experiment using MLfow
Given 28x28 pixels of handwritten digit, predict the number (0-9), hence this is multi-class classification problem.
- Accuracy
- F1 score
- Precision
- Recall
- Time taken for training
- Time taken for predicting
This notebook run using python 3.7. I use Anaconda to set the environment.
conda create --name new_environment python=3.7 # create new environment
conda activate new_environment # activate the env
pip install -r requirements.txt # install packages in requirements.txt