Expand description
Batch processing utilities for efficient embedding generation.
Structs§
- Batch
Processor - Batch processor for preparing text inputs for embedding models.
- Prepared
Batch - Prepared batch of tokenized inputs ready for ORT inference.
Functions§
- mean_
pooling - Apply mean pooling to ORT last-hidden-state output.
- normalize_
embeddings - Normalize embeddings to unit length (L2 normalization), in-place.