Enum EmbeddingInput 

pub enum EmbeddingInput {
    Single(String),
    Batch(Vec<String>),
    TokenIds(Vec<u32>),
    BatchTokenIds(Vec<Vec<u32>>),
}

Input accepted by the embeddings endpoint.

All four variants are deserialized from untagged JSON, so the format is inferred from the structure of the value supplied in the "input" field:

JSON value	Variant
"some text"	Single
["text one", "text two"]	Batch
[42, 1337]	TokenIds
[[42, 1337], [9, 99]]	BatchTokenIds

Variants

Single(String)

A single text string.

Batch(Vec<String>)

A batch of text strings.

TokenIds(Vec<u32>)

A single token-ID sequence (converted to a space-joined string).

BatchTokenIds(Vec<Vec<u32>>)

A batch of token-ID sequences.

Implementations

impl EmbeddingInput

pub fn as_strings(&self) -> Vec<String>

Convert all inputs to String form for embedding.

Token-ID sequences are rendered as space-separated decimal numbers so they can be passed through the text-based embedder.

pub fn len(&self) -> usize

Number of distinct inputs.

pub fn is_empty(&self) -> bool

Whether the input contains no items.

Trait Implementations

impl Debug for EmbeddingInput

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for EmbeddingInput

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.