ML Tutor : Learn Machine Learning while never leaving the conform of your Python IDE (Jupyter Notebook or Google Colab)
Project description
ML Tutor v1.0.3
ML Tutor is a Python library created to help people learn Machine Learning (ML).
Machine Learning (ML) is pretty hard and especially if you are just starting (been there done that)! I've created this library to help anybody interested in learning ML. ML Tutor provides visual training for every algorithm inside it so you can visualize what's happening with your data in real-time! Besides that, for every algorithm, there is a theory on how it works and interview questions.
Happy learning! ^_^
Use ML Tutor if you are looking to:
- Learn most popular Machine Learning algorithms directly from Jupyter Notebook or Google Colab
- Visualize what's happening with your data (Educational purpose only)
Usage
To demonstrate what you can do with ML Tutor, we will need a dataset. You can use your own dataset or some classic dataset (such as Iris).
Let's use the Iris dataset from the Sklearn library and split it into the training and testing subsets.
from sklearn.datasets import load_iris
from sklearn.model_selection import train_test_split
dataset = load_iris()
X_train, X_test, y_train, y_test = train_test_split(dataset.data, dataset.target, test_size=0.2)
Training/Prediction
Clustering
For the clustering example, let's import KMeans and demonstrate how to use it with the ML Tutor library.
Notice that you can train/test it just like any sklearn
algorithm.
Each algorithm has several arguments you can provide, but the unique one across all of them is visual_training
.
If you set this to True
, you will see the whole training process inside your IDE.
from ml_tutor.clustering.kmeans import KMeans
clf = KMeans(n_clusters=3, visual_training=True)
clf.fit(X_train)
Classification
For the classification example, let's use KNeighbourClassifier (KNN).
Because the KNN just stores data when you call the .fit()
function on it, the visualization part comes in the prediction time.
from ml_tutor.classification.knn import KNeighbourClassifier
clf = KNeighbourClassifier(n_neighbors=5, visual_training=True, number_of_visual_steps=2)
clf.fit(X_train, y_train)
predictions = clf.predict(X_test)
How an algorithm works? (theory)
Every algorithm has method .how_it_works()
which generates a blog post directly inside your IDE.
Every blog is written by somebody from the community, not myself, and in the end, they get a shout out for the great material.
from ml_tutor.classification.knn import KNeighbourClassifier
clf = KNeighbourClassifier()
clf.how_it_works()
What if I prefer video instead of reading tutorial?
I've got you covered! Just place video=True
inside the method how_it_works()
, and it will open a YouTube video for you.
from ml_tutor.classification.knn import KNeighbourClassifier
clf = KNeighbourClassifier()
clf.how_it_works(video=True)
Interview questions
If you call .interview_questions()
on any algorithm, it will generate resources with interview questions for the algorithm.
from ml_tutor.classification.knn import KNeighbourClassifier
clf = KNeighbourClassifier()
clf.interview_questions()
Sklearn code
Since this is the library for education and not for production, you'll need to learn how to use these algorithms with the battle-tested library sklearn
. Just call .sklearn_version()
on any algorithm, and it will generate code for you!
NOTE: For now, this method only works in Jupyter Notebook!
from ml_tutor.classification.knn import KNeighbourClassifier
clf = KNeighbourClassifier()
clf.sklearn_version()
Supported IDEs
For now this library is fully supported for Jupyter Notebook
and partially supported for Google Colab
(read Sklearn code
section for more details).
Installation
Installation with pip
You can install ML Tutor directly from the PyPi repository using pip
(or pip3
):
pip install ml-tutor
Manual installation
If you prefer to install it from source:
- Clone this repository
git clone https://github.com/lucko515/ml-tutor
- Go to the library folder and run
pip install .
TODO
If you want to contribute to the ML Tutor, here is what's on the TODO list:
-
.sklearn_version()
is not working in Google Colab - Logistic Regression visualization needs a re-do, currently it's not showing how the classification lines moves over time
- Interview questions should be added to each algorithm (take the
knn.py
as a reference) - Visualization export to
.gif
and/or.mp4
- Additional algorithms (e.g. NaiveBayes)
- Support for other IDEs (e.g. regular Python Shell)
Contact
Add me on LinkedIn.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
File details
Details for the file ml_tutor-1.0.3.tar.gz
.
File metadata
- Download URL: ml_tutor-1.0.3.tar.gz
- Upload date:
- Size: 26.3 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/3.2.0 pkginfo/1.5.0.1 requests/2.24.0 setuptools/50.3.0 requests-toolbelt/0.9.1 tqdm/4.49.0 CPython/3.7.9
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 585ecde93a2a214d9fc8acebe6a956e42cb9fc382ef9fc5f82d3d24a319abb03 |
|
MD5 | fedd2d2356df5faddd53ce949a83bdcd |
|
BLAKE2b-256 | bc9db9c0fb7574b684abcc704248001d52d968d3ce5ac9438688cf399e9b4d42 |