Wrapper around Fireworks API for large language models

Fireworks API is compatible to the OpenAI API with some limitations described in https://readme.fireworks.ai/docs/openai-compatibility.

To use, you should have the openai package installed and the FIREWORKS_API_KEY environment variable set.

Hierarchy

Constructors

  • Parameters

    • Optional fields: Partial<Omit<OpenAIInput, "openAIApiKey" | FireworksUnsupportedArgs>> & BaseLLMParams & {
          fireworksApiKey?: string;
      }

    Returns Fireworks

Properties

CallOptions: Partial<Omit<OpenAICallOptions, FireworksUnsupportedCallOptions>>
ParsedCallOptions: Omit<Partial<Omit<OpenAICallOptions, FireworksUnsupportedCallOptions>>, never>
batchSize: number

Batch size to use when passing multiple documents to generate

caller: AsyncCaller

The async caller should be used by subclasses to make any async calls, which will thus benefit from the concurrency and retry logic.

frequencyPenalty: number

Penalizes repeated tokens according to frequency

maxTokens: number

Maximum number of tokens to generate in the completion. -1 returns as many tokens as possible given the prompt and the model's maximum context size.

modelName: string

Model name to use

n: number

Number of completions to generate for each prompt

presencePenalty: number

Penalizes repeated tokens

streaming: boolean

Whether to stream the results or not. Enabling disables tokenUsage reporting

temperature: number

Sampling temperature to use

topP: number

Total probability mass of tokens to consider at each step

verbose: boolean

Whether to print out response text.

azureOpenAIApiDeploymentName?: string

Azure OpenAI API deployment name to use for completions when making requests to Azure OpenAI. This is the name of the deployment you created in the Azure portal. e.g. "my-openai-deployment" this will be used in the endpoint URL: https://{InstanceName}.openai.azure.com/openai/deployments/my-openai-deployment/

azureOpenAIApiInstanceName?: string

Azure OpenAI API instance name to use when making requests to Azure OpenAI. this is the name of the instance you created in the Azure portal. e.g. "my-openai-instance" this will be used in the endpoint URL: https://my-openai-instance.openai.azure.com/openai/deployments/{DeploymentName}/

azureOpenAIApiKey?: string

API key to use when making requests to Azure OpenAI.

azureOpenAIApiVersion?: string

API version to use when making requests to Azure OpenAI.

azureOpenAIBasePath?: string

Custom endpoint for Azure OpenAI API. This is useful in case you have a deployment in another region. e.g. setting this value to "https://westeurope.api.cognitive.microsoft.com/openai/deployments" will be result in the endpoint URL: https://westeurope.api.cognitive.microsoft.com/openai/deployments/{DeploymentName}/

bestOf?: number

Generates bestOf completions server side and returns the "best"

callbacks?: Callbacks
fireworksApiKey?: string
logitBias?: Record<string, number>

Dictionary used to adjust the probability of specific tokens being generated

metadata?: Record<string, unknown>
modelKwargs?: Record<string, any>

Holds any additional parameters that are valid to pass to openai.createCompletion that are not explicitly specified on this class.

openAIApiKey?: string

API key to use when making requests to OpenAI. Defaults to the value of OPENAI_API_KEY environment variable.

organization?: string
stop?: string[]

List of stop words to use when generating

tags?: string[]
timeout?: number

Timeout to use when making requests to OpenAI.

user?: string

Unique string identifier representing your end-user, which can help OpenAI to monitor and detect abuse.

Accessors

  • get callKeys(): string[]
  • Keys that the language model accepts as call options.

    Returns string[]

Methods

  • Default implementation of batch, which calls invoke N times. Subclasses should override this method if they can batch more efficiently.

    Parameters

    • inputs: BaseLanguageModelInput[]

      Array of inputs to each batch call.

    • Optional options: Partial<Partial<Omit<OpenAICallOptions, FireworksUnsupportedCallOptions>>> | Partial<Partial<Omit<OpenAICallOptions, FireworksUnsupportedCallOptions>>>[]

      Either a single call options object to apply to each batch call or an array for each call.

    • Optional batchOptions: RunnableBatchOptions & {
          returnExceptions?: false;
      }

    Returns Promise<string[]>

    An array of RunOutputs, or mixed RunOutputs and errors if batchOptions.returnExceptions is set

  • Parameters

    Returns Promise<(string | Error)[]>

  • Parameters

    Returns Promise<(string | Error)[]>

  • Convenience wrapper for generate that takes in a single string prompt and returns a single string output.

    Parameters

    • prompt: string
    • Optional options: string[] | Partial<Omit<OpenAICallOptions, FireworksUnsupportedCallOptions>>
    • Optional callbacks: Callbacks

    Returns Promise<string>

  • Calls the Fireworks API with retry logic in case of failures.

    Parameters

    • request: CompletionCreateParamsStreaming

      The request to send to the Fireworks API.

    • Optional options: OpenAICoreRequestOptions

      Optional configuration for the API call.

    Returns Promise<AsyncIterable<Completion>>

    The response from the Fireworks API.

  • Calls the Fireworks API with retry logic in case of failures.

    Parameters

    • request: CompletionCreateParamsNonStreaming

      The request to send to the Fireworks API.

    • Optional options: OpenAICoreRequestOptions

      Optional configuration for the API call.

    Returns Promise<Completion>

    The response from the Fireworks API.

  • Run the LLM on the given prompts and input, handling caching.

    Parameters

    • prompts: string[]
    • Optional options: string[] | Partial<Omit<OpenAICallOptions, FireworksUnsupportedCallOptions>>
    • Optional callbacks: Callbacks

    Returns Promise<LLMResult>

  • This method takes prompt values, options, and callbacks, and generates a result based on the prompts.

    Parameters

    • promptValues: BasePromptValue[]

      Prompt values for the LLM.

    • Optional options: string[] | Partial<Omit<OpenAICallOptions, FireworksUnsupportedCallOptions>>

      Options for the LLM call.

    • Optional callbacks: Callbacks

      Callbacks for the LLM call.

    Returns Promise<LLMResult>

    An LLMResult based on the prompts.

  • Parameters

    Returns Promise<number>

  • Get the identifying parameters for the model

    Returns Omit<CompletionCreateParams, "prompt"> & {
        model_name: string;
    } & ClientOptions

  • Get the parameters used to invoke the model

    Parameters

    • Optional options: Omit<Partial<Omit<OpenAICallOptions, FireworksUnsupportedCallOptions>>, never>

    Returns Omit<CompletionCreateParams, "prompt">

  • This method takes an input and options, and returns a string. It converts the input to a prompt value and generates a result based on the prompt.

    Parameters

    Returns Promise<string>

    A string result based on the prompt.

  • Create a new runnable sequence that runs each individual runnable in series, piping the output of one runnable into another runnable or runnable-like.

    Type Parameters

    • NewRunOutput

    Parameters

    • coerceable: RunnableLike<string, NewRunOutput>

      A runnable, function, or object whose values are functions or runnables.

    Returns RunnableSequence<BaseLanguageModelInput, Exclude<NewRunOutput, Error>>

    A new runnable sequence.

  • This method is similar to call, but it's used for making predictions based on the input text.

    Parameters

    • text: string

      Input text for the prediction.

    • Optional options: string[] | Partial<Omit<OpenAICallOptions, FireworksUnsupportedCallOptions>>

      Options for the LLM call.

    • Optional callbacks: Callbacks

      Callbacks for the LLM call.

    Returns Promise<string>

    A prediction based on the input text.

  • This method takes a list of messages, options, and callbacks, and returns a predicted message.

    Parameters

    • messages: BaseMessage[]

      A list of messages for the prediction.

    • Optional options: string[] | Partial<Omit<OpenAICallOptions, FireworksUnsupportedCallOptions>>

      Options for the LLM call.

    • Optional callbacks: Callbacks

      Callbacks for the LLM call.

    Returns Promise<BaseMessage>

    A predicted message based on the list of messages.

  • Stream output in chunks.

    Parameters

    Returns Promise<IterableReadableStream<string>>

    A readable stream that is also an iterable.

  • Stream all output from a runnable, as reported to the callback system. This includes all inner runs of LLMs, Retrievers, Tools, etc. Output is streamed as Log objects, which include a list of jsonpatch ops that describe how the state of the run has changed in each step, and the final state of the run. The jsonpatch ops can be applied in order to construct state.

    Parameters

    • input: BaseLanguageModelInput
    • Optional options: Partial<Partial<Omit<OpenAICallOptions, FireworksUnsupportedCallOptions>>>
    • Optional streamOptions: Omit<LogStreamCallbackHandlerInput, "autoClose">

    Returns AsyncGenerator<RunLogPatch, any, unknown>

  • Default implementation of transform, which buffers input and then calls stream. Subclasses should override this method if they can start producing output while input is still being generated.

    Parameters

    Returns AsyncGenerator<string, any, unknown>

Generated using TypeDoc