What is the ML Model Trainer multi-solver?

ML Model Trainer is a guided workflow that chains 4 individual PolySim solvers into one end-to-end analysis, piping each result into the next step.

Yes. Every step runs entirely in your browser using real numerics — no install, no account, no cloud cost. Custom or private solver packs are available as a paid service.

Data & Statistics Pack · Multi-Solver

ML Model Trainer

Train, cluster, reduce. This multi-solver chains 4 solvers into a single guided workflow — run each step in order and carry the result forward.

Workflow steps

or unlock everything with Pro →

Neural Net

open full solver →

Neural Network PlaygroundLive

2-layer MLP · live backprop

Controls

Presets

Hidden neurons

Learning rate

Speed1×

A tiny multilayer perceptron trained by real backpropagation. The background shows its learned decision boundary; dots are the training data, colored by true class.

▶ Run in Python

Data Inspector

Epoch0

Loss0.0000

Architecture2→8→1

Governing equation

a = tanh (W_{1} x + b_{1}), \overset{y}{^} = σ (W_{2} a + b_{2}), L = 0.0000

Reading this result: With 8 hidden units the net bends the boundary into curves — more units capture finer shapes like the spiral, but too many can overfit.

Runs locally in your browser — free forever. Scale to the cloud when reality gets heavy.

↓

Kmeans

open full solver →

k-Means ClusteringLive

Lloyd's algorithm · live

Controls

Speed1×

Presets

k (centroids)

True clusters

Lloyd's algorithm: assign each point to its nearest centroid, then move each centroid to the mean of its members. Repeat until stable. Try setting k different from the true cluster count.

▶ Run in Python

Data Inspector

Iteration0

Inertia0.00e+0

Governing equation

J = k = 1 \sum 4 x \in C_{k} \sum ∥ x - μ_{k} ∥^{2}, μ_{k} = \frac{1}{∣ C _{k} ∣} x \in C_{k} \sum x

Reading this result: k matches the 4 true clusters, so each centroid can settle onto one real group and inertia falls to a clean minimum.

Runs locally in your browser — free forever. Scale to the cloud when reality gets heavy.

↓

PCA

open full solver →

Principal Component AnalysisLive

finding the axes of variation

Controls

Presets

Correlation

Spread (x)

PCA finds the directions along which data varies most. The eigenvectors of the covariance matrix are the principal components (cyan = most variance, green = least), and their eigenvalues are the variances along each. Projecting onto the top components is the basis of dimensionality reduction.

▶ Run in Python

Data Inspector

Variance PC10.00

Variance PC20.00

PC1 explains0%

PC1 angle0°

Governing equation

C v = λ v, \frac{λ _{1}}{λ _{1} + λ _{2}} = \frac{0.00}{0.00} = 0.00

Reading this result: PC1 explains 0% of the variance here; the stronger the correlation between features, the more PCA concentrates information on that first component.

Runs locally in your browser — free forever. Scale to the cloud when reality gets heavy.

↓

Gradient Descent

open full solver →

Gradient Descent StudioLive

optimization on a loss landscape

Controls

Watch gradient descent (with momentum) roll downhill on different loss landscapes. Too high a learning rate overshoots.

Presets

Learning rate

Momentum

▶ Run in Python

Data Inspector

Surfaceripple

Step0/200

Optimizermomentum GD

Governing equation

x_{k + 1} = x_{k} - η \nabla f (x_{k}), η = 3

Reading this result: A learning rate of 3 is well matched to the curvature here: steps are large enough to make progress yet small enough to avoid overshoot, giving a smooth, roughly quadratic descent toward the minimum.

Runs locally in your browser — free forever. Scale to the cloud when reality gets heavy.

Need a private, branded version?

We build custom solvers and workflows for labs, firms, agencies, and courses — your parameters, your branding, your data.

Explore Custom Solver Sets →

More in the Data & Statistics pack

Classification Pipeline Clustering Workbench Feature Engineering Deep Net Playground Recommender System Anomaly Detection Time-Series Forecaster NLP Basics Kit

Frequently asked questions

What is the ML Model Trainer multi-solver?: ML Model Trainer is a guided workflow that chains 4 individual PolySim solvers into one end-to-end analysis, piping each result into the next step.
Is it free to use?: Yes. Every step runs entirely in your browser using real numerics — no install, no account, no cloud cost. Custom or private solver packs are available as a paid service.