Transforms text, images and videos into numerical embeddings that capture semantic meaning. This enables downstream tasks like semantic search, similarity comparison, and more.
Provide models that convert text into vector embeddings.
Offers OpenAI-based text embedding models.
modules
| Module Name | Module Description | Open Source Status | Cloud Status |
|---|---|---|---|
| ada-002 | Generates text embeddings using OpenAI’s Ada-002 model. | Alpha |
Available |
| embedding-3-large | Large-scale embedding model from OpenAI. | Alpha |
Available |
| embedding-3-small | Smaller variant for embedding text with lower resource usage. | Alpha |
Available |
Provides Cohere-based multilingual or English text embedding models.
modules
| Module Name | Module Description | Open Source Status | Cloud Status |
|---|---|---|---|
| multilingual-v3.0 | Cohere’s multilingual model (version 3.0) for text embeddings. | Backlog |
Available |
| multilingual-light-v3.0 | Lighter variant of the multilingual v3.0 model. | Backlog |
Available |
| english-v3.0 | Cohere’s English model (version 3.0) for text embeddings. | Backlog |
Available |
| english-light-v3.0 | Lightweight variant of the English v3.0 model. | Backlog |
Available |
| english-v2.0 | Cohere’s older English embedding model (version 2.0). | Backlog |
Available |
| english-light-v2.0 | Lightweight variant of the English v2.0 model. | Backlog |
Available |
Allows you to run open-source or self-deployed text embedding solutions.
modules
| Module Name | Module Description | Open Source Status | Cloud Status |
|---|---|---|---|
| text2vec-contextionary | A contextionary-based approach for generating embeddings locally. | Backlog |
Available |
| text2vec-transformers | Uses transformer models (e.g., via Hugging Face) for local text embedding. | Backlog |
Available |
user brings their own