Skip to content

Fireworks Context

The Fireworks Context uses models provided by the Fireworks API for the Action Engine's LLM and embedding models.

By default, it will use the open-source llama-v3p1-70b-instruct as the LLM & nomic-embed-text-v1.5 as the embedding model. You can change these models for other models provided by Fireworks.

Since there are no multi-modal LLMs available through the Fireworks API with LlamaIndex, we still use the default gpt-4o with OpenAI as the World Model mm_llm.

Pre-requisites

To use the Fireworks Context, you will first need to install the lavague FireworksContext package (as well as the lavague package if you haven't already):

pip install lavague-contexts-fireworks

API keys

You will need to set your Fireworks API & OpenAI API keys as FIREWORKS_API_KEY and OPENAI_API_KEY environment variables respectively. You can alternatively pass your API keys to the FireworksContext with the api_key andopenai_api_key parameters respectively.

For guidance on setting environment variables in your environment, see our FAQ.

End-to-end example

We can then import the FireworksContext from the lavague.contexts.fireworks package, initialize it and pass it to our Action Engine using the from_context() initialization method.

There is no need to pass it the World Model here since this context uses the default multi-modal (World Model) model.

from lavague.core import WorldModel, ActionEngine
from lavague.core.agents import WebAgent
from lavague.drivers.selenium import SeleniumDriver
from lavague.contexts.fireworks import FireworksContext

# Initialize Context
context = FireworksContext()

selenium_driver = SeleniumDriver()

# Build Action Engine and World Model from Context
action_engine = ActionEngine.from_context(context=context, driver=selenium_driver)
world_model = WorldModel()

# Build agent & run query
agent = WebAgent(world_model, action_engine)
agent.get("https://huggingface.co/")
agent.run("What is this week's top Space of the week?")

Customizing the Fireworks Context

You can change the Fireworks models used by the context by passing the name of the model you wish to use to the llm and embedding parameters respectively.

You should use the full name for your model as listed by Fireworks here.

Fireworks embedding models

Here is a list of currently available embedding models via the Fireworks API.

Model name Model size
nomic-ai/nomic-embed-text-v1.5 (recommended) 137M
nomic-ai/nomic-embed-text-v1 137M
WhereIsAI/UAE-Large-V1 335M
thenlper/gte-large 335M
thenlper/gte-base 109M 
from lavague.contexts.fireworks import FireworksContext

# Initialize Context
context = FireworksContext(llm="accounts/fireworks/models/code-llama-34b", embedding="nomic-ai/nomic-embed-text-v1")

While you can also change the multi-modal used by passing the name of the model as the mm_llm argument, this will need to be an OpenAI model since LlamaIndex does not yet support multi-modal models with Fireworks.