Struct TextClassifierDatasetBuilder

pub struct TextClassifierDatasetBuilder<'a, T: Class, E: Embedder> { /* private fields */ }

A builder for TextClassifier.

Example

// Create a dataset for the classifier
let bert = Bert::new().await?;
let mut dataset = TextClassifierDatasetBuilder::<MyClass, _>::new(&bert);
for question in ["What is the author's name?", "What is the author's age?"] {
    dataset.add(question, MyClass::Person).await?;
}
for question in [
    "What is the capital of France?",
    "What is the capital of England?",
] {
    dataset.add(question, MyClass::Thing).await?;
}

Implementations

impl<'a, T: Class, E: Embedder> TextClassifierDatasetBuilder<'a, T, E>

pub fn new(embedder: &'a E) -> Self

Creates a new TextClassifierDatasetBuilder.

pub async fn add( &mut self, text: impl ToString, class: T, ) -> Result<(), E::Error>

Adds a new example to the dataset.

pub async fn extend( &mut self, examples: impl IntoIterator<Item = (impl ToString, T)>, ) -> Result<(), E::Error>

Add many examples to the dataset at once. This may be faster than adding each example individually depending on the embedding model.

Example

// Create a dataset for the classifier
let bert = Bert::new().await?;
let mut dataset = TextClassifierDatasetBuilder::<MyClass, _>::new(&bert);
dataset
    .extend(
        ["What is the author's name?", "What is the author's age?"]
            .into_iter()
            .map(|q| (q, MyClass::Person)),
    )
    .await?;
dataset
    .extend(
        [
            "What is the capital of France?",
            "What is the capital of England?",
        ]
        .into_iter()
        .map(|q| (q, MyClass::Thing)),
    )
    .await?;

pub fn build(self, device: &Device) -> Result<ClassificationDataset>

Builds the dataset.