Implementation:Open compass VLMEvalKit StepAPI INT

Field	Value
source	VLMEvalKit
domain	Vision, API_Integration

Overview

StepAPI_INT provides a VLMEvalKit API adapter for Step AI vision-language models.

Description

StepAPI_INT inherits from BaseAPI and communicates with the Step AI chat completions endpoint at api.stepfun.com. It encodes images to base64 in WebP format for transmission, supports configurable temperature and max token parameters, and authenticates via the STEPAI_API_KEY environment variable with Bearer token authorization.

Usage

Use this adapter when evaluating Step AI models (such as step-1v-8k) through the StepFun API.

Code Reference

Source: vlmeval/api/stepai.py, Lines: L1-86
Import: from vlmeval.api.stepai import StepAPI_INT

Signature:

class StepAPI_INT(BaseAPI):
    def __init__(self, model='step-1v-8k', retry=10, key=None,
                 temperature=0, max_tokens=300, verbose=True,
                 system_prompt=None, **kwargs): ...
    def generate_inner(self, inputs, **kwargs): ...

I/O Contract

Direction	Description
Inputs	message — text/image/video content list; model-specific params via kwargs
Outputs	generate() returns str prediction; generate_inner() returns (int, str, str) tuple

Usage Examples

# Example instantiation
model = StepAPI_INT(model='step-1v-8k')
response = model.generate(message)

Related Pages

Page Connections

Double-click a node to navigate. Hold to expand connections.

Principle

Implementation

Heuristic

Environment