AIMV2-Large-Patch14-224-Lit Open-Source Vision Model - An Excellent and Practical Choice for Superior Multimodal Understanding Performance

Aimv2 Large Patch14 224 Lit

Developed by apple

AIMv2 is a series of vision models pretrained with multimodal autoregressive objectives, demonstrating outstanding performance across multiple multimodal understanding benchmarks.

Image-to-Text #Multimodal Autoregression #Zero-shot Classification #Open-vocabulary Detection

Downloads 222

Release Time : 11/20/2024

Model Overview

AIMv2 is pretrained with multimodal autoregressive objectives, showcasing robust performance in tasks such as image classification and object detection.

Model Features

Multimodal Autoregressive Pretraining

Utilizes innovative autoregressive objectives for pretraining to achieve superior multimodal understanding capabilities.

Exceptional Benchmark Performance

Outperforms OpenAI CLIP and SigLIP models in most multimodal understanding benchmarks.

Powerful Recognition Performance

The 3B version achieves 89.5% accuracy on ImageNet when using a frozen backbone network.

Broad Application Capabilities

Surpasses DINOv2 in open-vocabulary object detection and referring expression comprehension tasks.

Model Capabilities

Zero-shot image classification

Multimodal understanding

Open-vocabulary object detection

Referring expression comprehension

Use Cases

Computer Vision

Image Classification

Classify and recognize image content

89.5% accuracy on ImageNet

Object Detection

Detect specific objects in images

Outperforms DINOv2 model

Multimodal Applications

Image-Text Matching

Understand the relationship between images and text descriptions

Surpasses CLIP and SigLIP models

🚀 Transformers

This project introduces the AIMv2 family of vision models pre - trained with a multimodal autoregressive objective, which shows excellent performance in multiple multimodal understanding benchmarks.

🚀 Quick Start

We introduce the AIMv2 family of vision models pre - trained with a multimodal autoregressive objective. AIMv2 pre - training is simple and straightforward to train and to scale effectively. Some AIMv2 highlights include:

Outperforms OAI CLIP and SigLIP on the majority of multimodal understanding benchmarks.
Outperforms DINOv2 on open - vocabulary object detection and referring expression comprehension.
Exhibits strong recognition performance with AIMv2 - 3B achieving 89.5% on ImageNet using a frozen trunk.

💻 Usage Examples

Basic Usage

import requests
from PIL import Image
from transformers import AutoProcessor, AutoModel

url = "http://images.cocodataset.org/val2017/000000039769.jpg"
image = Image.open(requests.get(url, stream=True).raw)
text = ["Picture of a dog.", "Picture of a cat.", "Picture of a horse."]

processor = AutoProcessor.from_pretrained(
    "apple/aimv2-large-patch14-224-lit",
)
model = AutoModel.from_pretrained(
    "apple/aimv2-large-patch14-224-lit",
    trust_remote_code=True,
)

inputs = processor(
    images=image,
    text=text,
    add_special_tokens=True,
    truncation=True,
    padding=True,
    return_tensors="pt",
)
outputs = model(**inputs)
probs = outputs.logits_per_image.softmax(dim=-1)

Advanced Usage

Under construction.

📄 License

The project uses the apple - amlr license.

📚 Documentation

[AIMv2 Paper] [BibTeX]

📦 Citation

If you find our work useful, please consider citing us as:

@misc{fini2024multimodalautoregressivepretraininglarge,
  author      = {Fini, Enrico and Shukor, Mustafa and Li, Xiujun and Dufter, Philipp and Klein, Michal and Haldimann, David and Aitharaju, Sai and da Costa, Victor Guilherme Turrisi and Béthune, Louis and Gan, Zhe and Toshev, Alexander T and Eichner, Marcin and Nabi, Moin and Yang, Yinfei and Susskind, Joshua M. and El - Nouby, Alaaeldin},
  url         = {https://arxiv.org/abs/2411.14402},
  eprint      = {2411.14402},
  eprintclass = {cs.CV},
  eprinttype  = {arXiv},
  title       = {Multimodal Autoregressive Pre - training of Large Vision Encoders},
  year        = {2024},
}

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご