Working with Fireworks AI in Pixeltable

    

This documentation page is also available as an interactive notebook. You can launch the notebook in Kaggle or Colab, or download it for use with an IDE or local Jupyter installation, by clicking one of the above links.

Pixeltable’s Fireworks integration enables you to access LLMs hosted on the Fireworks platform.

Prerequisites

• A Fireworks account with an API key (https://fireworks.ai/api-keys)

Important Notes

• Fireworks usage may incur costs based on your Fireworks plan.
• Be mindful of sensitive data and consider security measures when integrating with external services.

First you’ll need to install required libraries and enter a Fireworks API key.

%pip install -qU pixeltable fireworks-ai

import os
import getpass

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

Now let’s create a Pixeltable directory to hold the tables for our demo.

import pixeltable as pxt

# Remove the 'fireworks_demo' directory and its contents, if it exists
pxt.drop_dir('fireworks_demo', force=True)
pxt.create_dir('fireworks_demo')

Connected to Pixeltable database at: postgresql+psycopg://postgres:@/pixeltable?host=/Users/asiegel/.pixeltable/pgdata
Created directory ‘fireworks_demo’.
<pixeltable.catalog.dir.Dir at 0x301799de0>

Completions

Create a Table: In Pixeltable, create a table with columns to represent your input data and the columns where you want to store the results from Fireworks.

from pixeltable.functions.fireworks import chat_completions

# Create a table in Pixeltable and pick a model hosted on Fireworks with some parameters

t = pxt.create_table('fireworks_demo.chat', {'input': pxt.String})

messages = [{'role': 'user', 'content': t.input}]
t.add_computed_column(output=chat_completions(
    messages=messages,
    model='accounts/fireworks/models/mixtral-8x22b-instruct',
    model_kwargs={
        # Optional dict with parameters for the Fireworks API
        'max_tokens': 300,
        'top_k': 40,
        'top_p': 0.9,
        'temperature': 0.7
    }
))

Created table `chat`.
Added 0 column values with 0 errors.
UpdateStatus(num_rows=0, num_computed_values=0, num_excs=0, updated_cols=[], cols_with_excs=[])

# Parse the bot_response into a new column
t.add_computed_column(response=t.output.choices[0].message.content)

Added 0 column values with 0 errors.
UpdateStatus(num_rows=0, num_computed_values=0, num_excs=0, updated_cols=[], cols_with_excs=[])

# Start a conversation
t.insert([{'input': 'Can you tell me who was President of the US in 1961?'}])
t.select(t.input, t.response).show()

Inserting rows into `chat`: 1 rows [00:00, 85.10 rows/s]
Inserted 1 row with 0 errors.

Learn More

To learn more about advanced techniques like RAG operations in Pixeltable, check out the RAG Operations in Pixeltable tutorial.

If you have any questions, don’t hesitate to reach out.