Expand description
Batch text encoding pipeline with chunking and pooling strategies.
Provides deterministic text-to-embedding conversion with configurable pooling, normalization, and similarity computation — all without external ML dependencies.
Structs§
- Batch
Encoder - Batch text encoder: tokenises, embeds, pools, and normalises text.
- Encoded
Batch - The output of encoding a batch of texts.
- Encoding
Config - Configuration for the batch encoder.
- Tokenized
Text - A tokenised representation of a single text string.
Enums§
- Pooling
Strategy - Pooling strategy for aggregating token-level embeddings.