Machine Learning Calculators
Free, interactive calculators for the mathematics that powers modern AI — linear algebra, statistics, calculus, and the models themselves. Every tool shows step-by-step working and the matching NumPy code. No sign-up, no paywall.
Browse by topic
Linear Algebra
Matrix multiplication, determinant, transpose, dot & cross product.
Open the hub → 📊Statistics & Probability
Percentiles, quartiles, combinations, sampling odds and expected value.
Open the hub → ∫Calculus
Derivatives, integrals, gradients, limits, and Taylor series.
🧠Deep Learning
Layer shapes, parameter counts, GPU memory, training steps, learning rate and loss.
Open the hub → 🎯Model Evaluation
Confusion matrix, F1 and F-beta, ROC AUC, entropy and Gini impurity.
Open the hub →Popular calculators
Start with the tools at the heart of every machine learning model:
- Matrix Multiplication Calculator — the operation inside every neural-network layer.
- Matrix Determinant Calculator — tells you whether a matrix is invertible.
- Matrix Transpose Calculator — flips rows and columns, everywhere in gradient math.
- Dot Product Calculator — measures how aligned two vectors are, with the angle.
- Cross Product Calculator — the perpendicular vector for 3D geometry.
- Conv2d Output Size Calculator — the shape a convolution or pooling layer returns, before you run the code.
- LLM VRAM Calculator — whether a language model fits on your GPU, split into weights and KV cache.
- F1 Score Calculator — F-beta plus macro, micro and weighted F1 across classes.
- Percentile Calculator — where a value sits in your data, by the method NumPy uses or by nearest rank.
- Expected Value Calculator — the probability-weighted average that every loss function estimates.
Why learn the math behind machine learning?
You can train a model without knowing the mathematics — until something breaks. When a network will not converge, a metric looks wrong, or a tutorial glosses over a step, it is the underlying linear algebra, statistics, and calculus that let you diagnose and fix it. These calculators make that math tangible: enter real numbers, see the result and the working, then reproduce it in a line of NumPy. Understanding beats memorising, and doing beats reading.
How these calculators are different
Most online calculators just print a number. Ours are teaching tools. Each one reveals the intermediate steps — the row-by-column products behind a matrix multiplication, the cofactor expansion behind a determinant — so the method sticks. Each is paired with a written guide that explains what the operation means, where it appears in machine learning, and the exact Python function to reach for in practice. And because they are built on a deliberately lightweight foundation, they load fast on any device.
Who makes these calculators, and how are they checked?
MLCalculators is built and maintained by Jamal Hussein, an MSc student in Artificial Intelligence. Each calculator’s maths is run against test cases with known answers, worked from the exact formula or produced by NumPy and SciPy, before the page is published. The written explanations are drafted with AI assistance and their worked examples are checked against the tested calculator.
The numbers you type stay in your browser; the site has no accounts, cookies or tracking, as the privacy policy sets out.
Frequently asked questions
Are the calculators really free?
Yes — every tool on ML Calculators is completely free, requires no account, and runs entirely in your browser.
Do I need to know machine learning to use them?
Not at all. The calculators work for any student of mathematics; the guides simply add machine-learning context for those who want it.
Which topics are covered?
Linear algebra, regression, vectors, core machine-learning functions and a first set of statistics calculators are live, with calculus still to come. Start at the linear algebra hub.