Fireworks AI

Introduction

Fireworks AI offers a rapid, cost-effective, and adaptable solution for generative artificial intelligence, enabling product developers to execute, refine, and distribute LLMs efficiently.

Deploying your model

Fireworks offers the capability to deploy various Chat Models(LLM) under its umbrella.

• Environment Variable :

  Below value(s) are to be added in .env file, present at builder/partnerproduct/.

  FIREWORKS_API_KEY : <check_references_below>

Chat Model

Refer to their documentation to understand the latest offerings, with feature and cost comparisons.

Usage with MAAP

To use Fireworks model with MAAP framework, you would need to feed below values.

• Config File :

  Provided below are the values required to be added in config.yaml file in embedding section.

  llms:
      class_name: Fireworks
      model_name: <check_references_below>
      max_tokens: <integer_value>
      temperature: <integer_value>

Embedding Model

You can follow the same steps as above to deploy the embedding model as well. The process is documented here.

Usage with MAAP

To use FireworksAI embedding with MAAP framework, you would need to feed below values.

• Config File :

  Provided below are the values required to be added in config.yaml file in embedding section. The class_name should be appropriate to the embedding model being used.

  embedding:
      class_name: Fireworks
      model_name: <check_references_below> 

Deploying your model using the LlamaIndex framework

Chat Model

MAAP now provides the option to choose if you want to use LlamaIndex as your main framework to deploy your LLM models.

This can be done by adding the 'framework' configuration to the config.yaml file

• Config File

  llms:
    class_name: Fireworks
    model_name: <check_references_below>
    framework: 'llamaindex'

Embedding Model

MAAP now provides the option to choose if you want to use LlamaIndex as your main framework to deploy your embeddings.

This can be done by adding the 'framework' configuration to the config.yaml file

• Config File

  embedding:
    class_name: Fireworks
    model_name: <check_references_below>
    framework: 'llamaindex'
  Another important consideration to have is that because of the way that LlamaIndex implements its embedding models, the environment variables that you are using cannot contain any parameters related to AzureOpenAI. This is because LlamaIndex automatically detects if you have any Azure environment variables and uses their endpoints if they are set.

References

Provided below are the instructions on how to procure the right values for building your MAAP framework.

• Chat Model Name

  You can pick any model from the updated LLM model list given in Fireworks documentation.

• Embedding Model Name

  You can pick any model from the updated embedding model list given in Fireworks documentation.

• API Key

  You will need to sign up and retrieve a Fireworks API Key.