Qwen/Qwen3-VL-Embedding-2B

Note

This is a remotely-sourced model from the qwen3vl_embeddings plugin, maintained by the community. It is not part of FiftyOne core and may have special installation requirements. Please review the plugin documentation and license before use.

Qwen3-VL-Embedding Specifically designed for multimodal information retrieval and cross-modal understanding, this suite accepts diverse inputs including text, images, screenshots, and videos, as well as inputs containing a mixture of these modalities.

Details

• Model name: Qwen/Qwen3-VL-Embedding-2B

• Model source: https://huggingface.co/Qwen/Qwen3-VL-4B-Instruct

• Model author: Qwen

• Model license: Apache-2.0

• Exposes embeddings? yes

• Tags: classification, torch, zero-shot, video, embeddings

Requirements

• Packages: huggingface-hub, transformers, torch, torchvision, qwen-vl-utils, decord

• CPU support

  • yes

• GPU support

  • yes

Example usage

import fiftyone as fo
import fiftyone.zoo as foz

foz.register_zoo_model_source("https://github.com/harpreetsahota204/qwen3vl_embeddings")

dataset = foz.load_zoo_dataset(
    "coco-2017",
    split="validation",
    dataset_name=fo.get_default_dataset_name(),
    max_samples=50,
    shuffle=True,
)

model = foz.load_zoo_model("Qwen/Qwen3-VL-Embedding-2B")

dataset.apply_model(model, label_field="predictions")

session = fo.launch_app(dataset)