Enables calls to the Google Cloud's Vertex AI API to access Large Language Models in a chat-like fashion.

This entrypoint and class are intended to be used in web environments like Edge functions where you do not have access to the file system. It supports passing service account credentials directly as a "GOOGLE_VERTEX_AI_WEB_CREDENTIALS" environment variable or directly as "authOptions.credentials".

const model = new ChatGoogleVertexAI({
temperature: 0.7,
});
const result = await model.invoke(
"How do I implement a binary search algorithm in Python?",
);

Hierarchy

  • BaseChatGoogleVertexAI<WebGoogleAuthOptions>
    • ChatGoogleVertexAI

Constructors

Properties

connection: GoogleVertexAILLMConnection<BaseLanguageModelCallOptions, GoogleVertexAIChatInstance, GoogleVertexAIChatPrediction, WebGoogleAuthOptions>
examples: ChatExample[] = []

Help the model understand what an appropriate response is

maxOutputTokens: number = 1024

Maximum number of tokens to generate in the completion.

model: string = "chat-bison"

Model to use

streamedConnection: GoogleVertexAILLMConnection<BaseLanguageModelCallOptions, GoogleVertexAIChatInstance, GoogleVertexAIChatPrediction, WebGoogleAuthOptions>
temperature: number = 0.2

Sampling temperature to use

topK: number = 40

Top-k changes how the model selects tokens for output.

A top-k of 1 means the selected token is the most probable among all tokens in the model’s vocabulary (also called greedy decoding), while a top-k of 3 means that the next token is selected from among the 3 most probable tokens (using temperature).

topP: number = 0.8

Top-p changes how the model selects tokens for output.

Tokens are selected from most probable to least until the sum of their probabilities equals the top-p value.

For example, if tokens A, B, and C have a probability of .3, .2, and .1 and the top-p value is .5, then the model will select either A or B as the next token (using temperature).

Methods