Dense Vector Embeddings & Feature Extraction

4 repos

Model weights, implementations, and tooling for generating dense vector embeddings from text and multimodal inputs using transformer-based architectures. This area covers embedding models optimized for semantic search, retrieval-augmented generation, and feature extraction tasks—including specialized variants like Llama-2, Llama-3, and Mistral-based models adapted for embedding generation, as well as multimodal embedding models. Repos are published in safetensors format and expose transformer-compatible endpoints for inference.

Related clusters