Embeddings NVIDIAInstall

Summary

Generate text embeddings using a selected NVIDIA model via the NVIDIA NIM API, supporting configurable encoding and truncation options.

Inputs

• Model Name or ID (required) — Select a pretrained NVIDIA embedding model from a dynamically loaded list or specify a custom model ID to use for embedding generation.
• Encoding Format — Choose the encoding format for the output embeddings: 'float' for numeric arrays or 'base64' for encoded strings.
• Truncate — Select how to truncate input text when it exceeds the model's limits: no truncation, truncate at the start, or at the end.

Output shape

An embedding vector or list of embedding vectors corresponding to the input text, depending on usage context.

The output embeddings are provided as arrays or base64 strings according to encoding format. The operation returns an object exposing methods for embedding a query (single text) or documents (multiple texts). Errors occur if input text is empty or API responses are unexpected.