paraphrase-multilingual-minilm-l12-v2

This is a Cog wrapper for the sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2 model, published by AudioScrape.

It generates 384-dimensional multilingual sentence embeddings optimized for semantic search across more than 50 languages.

This model is used by audioscrape.com to power fast and accurate audio transcription search in multiple languages.

✨ Example

cog predict -i inputs='["hello world", "hola mundo", "bonjour le monde"]'

📘 Model Details

Name: paraphrase-multilingual-MiniLM-L12-v2

Source: Hugging Face

Embedding size: 384 dimensions

Supported languages: 50+

Use case: Multilingual sentence embeddings for semantic search

🔗 Attribution & License

Original model by SBERT.net

Licensed under Apache 2.0

This wrapper is maintained by AudioScrape