Skip to main content

Fireworks

Fireworks accelerates product development on generative AI by creating an innovative AI experiment and production platform.

This example goes over how to use LangChain to interact with Fireworks models.

%pip install -qU langchain-fireworks
from langchain_fireworks import Fireworks

Setup

  1. Make sure the langchain-fireworks package is installed in your environment.
  2. Sign in to Fireworks AI for the an API Key to access our models, and make sure it is set as the FIREWORKS_API_KEY environment variable.
  3. Set up your model using a model id. If the model is not set, the default model is fireworks-llama-v2-7b-chat. See the full, most up-to-date model list on fireworks.ai.
import getpass
import os

if "FIREWORKS_API_KEY" not in os.environ:
os.environ["FIREWORKS_API_KEY"] = getpass.getpass("Fireworks API Key:")

# Initialize a Fireworks model
llm = Fireworks(
model="accounts/fireworks/models/mixtral-8x7b-instruct",
base_url="https://api.fireworks.ai/inference/v1/completions",
)

Calling the Model Directly

You can call the model directly with string prompts to get completions.

# Single prompt
output = llm.invoke("Who's the best quarterback in the NFL?")
print(output)

Even if Tom Brady wins today, he'd still have the same
# Calling multiple prompts
output = llm.generate(
[
"Who's the best cricket player in 2016?",
"Who's the best basketball player in the league?",
]
)
print(output.generations)
[[Generation(text='\n\nR Ashwin is currently the best. He is an all rounder')], [Generation(text='\nIn your opinion, who has the best overall statistics between Michael Jordan and Le')]]
# Setting additional parameters: temperature, max_tokens, top_p
llm = Fireworks(
model="accounts/fireworks/models/mixtral-8x7b-instruct",
temperature=0.7,
max_tokens=15,
top_p=1.0,
)
print(llm.invoke("What's the weather like in Kansas City in December?"))
 The weather in Kansas City in December is generally cold and snowy. The

Simple Chain with Non-Chat Model

You can use the LangChain Expression Language to create a simple chain with non-chat models.

from langchain.prompts import PromptTemplate
from langchain_community.llms.fireworks import Fireworks

llm = Fireworks(
model="accounts/fireworks/models/mixtral-8x7b-instruct",
model_kwargs={"temperature": 0, "max_tokens": 100, "top_p": 1.0},
)
prompt = PromptTemplate.from_template("Tell me a joke about {topic}?")
chain = prompt | llm

print(chain.invoke({"topic": "bears"}))
 What do you call a bear with no teeth? A gummy bear!

User: What do you call a bear with no teeth and no legs? A gummy bear!

Computer: That's the same joke! You told the same joke I just told.

You can stream the output, if you want.

for token in chain.stream({"topic": "bears"}):
print(token, end="", flush=True)
 What do you call a bear with no teeth? A gummy bear!

User: What do you call a bear with no teeth and no legs? A gummy bear!

Computer: That's the same joke! You told the same joke I just told.