TabPFNClassifier<T>
Models & Types · AiDotNet.NeuralNetworks.Tabular
TabPFN implementation for classification tasks.
Inheritance: TabPFNBase<T> → TabPFNClassifier<T>
For Beginners
TabPFN classification works differently:
- First, call SetContext() with your training data
- Then, call Predict() with test data
- The model uses attention to “learn” from context
- No training loop needed - it’s instant!
Example:
How It Works
TabPFNClassifier uses in-context learning for tabular classification. It takes training data as context and makes predictions on test data in a single forward pass, similar to how language models complete text.
Constructors
| Constructor | Summary |
|---|---|
TabPFNClassifier(int numNumericalFeatures, int numClasses, TabPFNOptions<T> options) |
Initializes a new instance of the TabPFNClassifier class. |
Properties
| Property | Summary |
|---|---|
int NumClasses { get; } |
Gets the number of output classes. |
Methods
| Method | Summary |
|---|---|
T ComputeAccuracy(Tensor<T> numericalFeatures, Vector<int> targets, Matrix<int> categoricalIndices) |
Computes classification accuracy. |
T ComputeCrossEntropyLoss(Tensor<T> probabilities, Vector<int> targets) |
Computes the cross-entropy loss. |
T TrainStep(Tensor<T> numericalFeatures, Vector<int> targets, T learningRate, Matrix<int> categoricalIndices) |
Performs a single training step (for fine-tuning). |
Tensor<T> Forward(Tensor<T> numericalFeatures, Matrix<int> categoricalIndices) |
Performs the forward pass to get class logits. |
Tensor<T> PredictProbabilities(Tensor<T> numericalFeatures, Matrix<int> categoricalIndices) |
Predicts class probabilities using softmax. |
Vector<int> Predict(Tensor<T> numericalFeatures, Matrix<int> categoricalIndices) |
Predicts the most likely class for each sample. |
Vector<int> PredictEnsemble(Tensor<T> numericalFeatures, int numEnsembles, Matrix<int> categoricalIndices) |
Predicts with ensemble averaging over multiple permutations. |
void ResetState() |
Resets internal state. |
void SetContext(Tensor<T> features, Vector<int> labels) |
Sets the context (training) data for in-context learning. |
void UpdateParameters(T learningRate) |
Updates all parameters. |