Unlocking the Power of Fireworks AI with Nomic-Embed-Text-V1.5

Fireworks AI has integrated the advanced Nomic-Embed-Text-V1.5 model to offer users unparalleled text embedding capabilities. This integration brings a host of benefits, including superior performance in embedding tasks and flexible cost options.

Nomic-Embed-Text-V1.5 Model

The Nomic-Embed-Text-V1.5 model stands out as the first text embedding model with a context length of 8192. It outperforms OpenAI's Ada-002 and text-embedding-3-small models on both short and long context tasks. The model leverages a Transformer architecture and is trained using a self-supervised MLM objective, followed by contrastive training on web-scale unsupervised data and fine-tuning on a curated corpus of paired data.

Released under an Apache-2 license, this model ensures full auditability, providing access to both model weights and training code.

Integration with Fireworks AI

Fireworks AI now supports the Nomic-Embed-Text-V1.5 model, allowing users to generate embeddings from text inputs. Notably, the model features a variable embedding size, which can be adjusted based on cost sensitivity.

To get started, create a Fireworks account, obtain an API key, and install the necessary integration packages, such as langchain-fireworks for LangChain integration.

Usage

Generating embeddings with Fireworks AI is straightforward. Use the FireworksEmbeddings class from the langchain-fireworks package. Initialize the model with the nomic-ai/nomic-embed-text-v1.5 model name and use it to embed documents or queries.

from langchain_fireworks import FireworksEmbeddings

embeddings = FireworksEmbeddings(model="nomic-ai/nomic-embed-text-v1.5")

You can then use the embed_documents or embed_query methods to create embeddings for your texts.

Benefits and Features

The Nomic-Embed-Text-V1.5 model is optimized for retrieval-augmented generation (RAG) and semantic search. It encodes semantic information into low-dimensional vectors and is compatible with various vector databases like Chroma, enabling efficient indexing and retrieval of documents.

Setup and Credentials

To use Fireworks AI with the Nomic-Embed-Text-V1.5 model, set your Fireworks API key as an environment variable:

import os
os.environ["FIREWORKS_API_KEY"] = getpass.getpass("Enter your Fireworks API key: ")

Then, install the necessary packages and initialize the model as described above.