spring-projects/spring-ai

Support for VertexAI custom models (Endpoints)

Open

#654 opened on Apr 29, 2024

View on GitHub
 (2 comments) (0 reactions) (1 assignee)Java (2,617 forks)auto 404
gcphelp wantedintegration testingvertex

Repository metrics

Stars
 (8,872 stars)
PR merge metrics
 (Avg merge 2d 1h) (50 merged PRs in 30d)

Description

Expected Behavior

I would like to be able to use my Custom Model (e.g. fine-tuned foundation model) deployed on GCP Endpoint with Spring AI library.

API for Endpoints: https://cloud.google.com/vertex-ai/docs/reference/rest/v1/projects.locations.endpoints Example for predict: https://us-central1-aiplatform.googleapis.com/v1/projects/49...78/locations/us-central1/endpoints/433...736:predict {"instances":[{"content":"Hello"}],"parameters":{"temperature":0.2,"maxOutputTokens":1024,"topP":0.8,"topK":40,"candidateCount":1}}

Current Behavior

Currently in VertexAI PaLM2 Chat and VertexAI Gemini Chat you can only use Foundation models: https://cloud.google.com/vertex-ai/generative-ai/docs/model-reference/text-chat

Context

There is a similar issue for langchain4j: https://github.com/langchain4j/langchain4j/issues/440

Contributor guide

Research direction
Review the existing VertexAI PaLM2 and Gemini chat implementations in Spring AI. Understand the API for Vertex AI custom model endpoints and identify the necessary changes to support fine tuned models. Look at how the API request is structured for endpoints vs. foundation models.
Tech stack
java
Domain
aibackendcloud
Issue type
Feature
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
3
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
1-2 days
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
Active
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
Mostly clear
Prerequisites
JavaSpring AIVertex AI
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
40