Implementation:Open compass VLMEvalKit VITAQwen2

Field	Value
source	VLMEvalKit
domain	Vision, Model_Architecture

Overview

VLM adapter for the VITA model enabling benchmark evaluation in VLMEvalKit.

Description

VITAQwen2 inherits from BaseModel and wraps the VITA model for use within the VLMEvalKit evaluation framework. It initializes the model and tokenizer/processor from a HuggingFace model path (default: VITA/vita) and provides the generate_inner method for inference. Also includes the VITA adapter class. VITAQwen2 wraps the VITA model with Qwen2.5 instruction template.

Usage

Register in vlmeval/config.py via supported_VLM and invoke through the standard evaluation pipeline.

Code Reference

Source: vlmeval/vlm/vita.py, Lines: L1-425
Import: from vlmeval.vlm.vita import VITAQwen2

Signature:

class VITAQwen2(BaseModel):
    INSTALL_REQ = True
    INTERLEAVE = True
    def __init__(self, model_path='VITA/vita', **kwargs): ...
    def generate_inner(self, message, dataset=None): ...

I/O Contract

Direction	Description
Inputs	message — list of dicts with type (text/image) and value; dataset — optional dataset name for custom prompting
Outputs	generate_inner() returns str (model response text)

Usage Examples

from vlmeval.vlm.vita import VITAQwen2
model = VITAQwen2(model_path='path/to/model')
response = model.generate_inner(message)

Related Pages

Page Connections

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

Principle

Implementation

Heuristic

Environment