Our take
Inactiveno commits in over a year- Activity
- 19
- Community
- 19
- Issues
- 21
- Pull requests
- 20
+376 stars in the last 30 days
This is a broad, approachable learning resource, but its inactive maintenance and highly concentrated contribution history make it a poor default for production dependencies.
Good fit if
- You want to study or experiment with readable implementations across several machine learning areas.
- You are comfortable with Python and want examples whose stated priority is transparency rather than computational efficiency.
- MIT licensing suits your intended use.
Look elsewhere if
- You need actively maintained code: the project has had no recent commits or releases, and recent issue and pull-request activity shows no recorded closures or merges.
- You need a project with distributed maintainer ownership: most contributions come from one person.
- You require optimized, production-oriented algorithms rather than educational implementations.
All health signals
| Last commit | 2023-10-15 (3 years ago) |
|---|---|
| Commits, last 90 days | 0 |
| Releases, last 12 months | 0 |
| Contributors | 6 (top contributor: 99% of commits) |
| Issues closed, last 90 days | 0 |
| Pull requests merged, last 90 days | 0 |
| Project age | 9 years |
Checked on 2026-10-08 with the GitHub API.
Project history
Daily GitHub snapshots since Oct 7, 2026. History covers stars and forks; health signals are not tracked historically.
Stars
33,383Forks
5,546Change since Nov 17, 2026: +9 stars·+1 forks
Overview
ML-From-Scratch is a Python collection of machine learning algorithms implemented with NumPy. Its main value is educational: implementations expose how models work rather than prioritizing speed or production-scale efficiency.
The range spans classical supervised and unsupervised methods, deep learning components, and a deep Q-network. It suits readers who want to inspect or experiment with algorithm implementations; it is less suited to teams seeking a maintained, performance-focused ML platform.
Key Features
- Supervised learning implementations include regression, decision trees, random forests, support vector machines, and boosting methods.
- Unsupervised methods include clustering, dimensionality reduction, association analysis, and generative models.
- Neural-network code includes layers such as convolution, pooling, batch normalization, and recurrent layers.
- Example programs demonstrate tasks such as digit classification, handwritten-digit generation, and CartPole with a deep Q-network.
- Genetic algorithms and neuroevolution provide examples of evolutionary approaches to optimization and neural networks.
- Installation uses the repository's Python setup script, and the project is licensed under MIT.
Use Cases
- Students learning how regression, classification, or clustering algorithms are structured can read and modify the NumPy implementations.
- Instructors can use the examples to demonstrate model behavior and training concepts without treating the code as a high-performance library.
- Developers exploring a specific algorithm can use the implementations as a starting point for experiments or comparison with other approaches.
- Readers studying neural-network internals can inspect the included layers and example CNN, RNN, and MLP models.
What you need
Detected in the repository
- Python (from setup.py)
License in plain words
MITpermissive
- Commercial use: yes
- Modify and redistribute: yes
- You must keep: the copyright and license notice
- Share your changes: no
A summary, not legal advice: the LICENSE file is what applies.
Getting Started
git clone https://github.com/eriklindernoren/ML-From-Scratch
cd ML-From-Scratch
python setup.py install
See the README for example commands and the full implementation list.
Considerations
The repository is inactive, with no recent commits or releases, and its contribution history is heavily concentrated in one person. The supplied project data also reports no CI or test suite, so adopters should independently check behavior and compatibility before relying on it. The README explicitly prioritizes accessibility over computational efficiency. Python is the stated runtime; the supplied project information does not specify GPU or database requirements.
Found this useful?
Share it with someone who would like ML-From-Scratch.