jkrukowski/swift-embeddings

Run embedding models locally in Swift using MLTensor.

52
/ 100
Established

Supports multiple transformer architectures (BERT, RoBERTa, XLM-RoBERTa, ModernBERT, CLIP) and lightweight embedding models (Word2Vec, Model2Vec) with configurable weight loading and batch processing via the MLTensor framework. Models load directly from Hugging Face with optional key transformations to handle architecture-specific weight naming conventions. Includes a command-line interface and integrates with MLTensorUtils for operations like cosine distance computation on encoded tensors.

139 stars.

No Package No Dependents
Maintenance 10 / 25
Adoption 10 / 25
Maturity 16 / 25
Community 16 / 25

How are scores calculated?

Stars

139

Forks

19

Language

Swift

License

MIT

Last pushed

Feb 07, 2026

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/embeddings/jkrukowski/swift-embeddings"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.