E

Eagle2 2B

Developed by nvidia
Eagle2 is a high-performance vision-language model family introduced by NVIDIA, focusing on enhancing the performance of open-source vision-language models through data strategies and training approaches. Eagle2-2B is the lightweight model in this series, achieving outstanding efficiency and speed while maintaining robust performance.
Downloads 667
Release Time : 1/10/2025

Model Overview

Eagle2-2B is a multimodal model integrating vision and language capabilities, capable of processing image, text, and video inputs to perform various vision-language tasks.

Model Features

Efficient and Lightweight
Achieves an excellent balance of performance and speed at the 2B parameter scale
Multimodal Processing
Supports comprehensive processing capabilities for image, text, and video inputs
Long Context Support
Supports context lengths of up to 16K tokens
High-Performance Benchmark
Delivers outstanding performance across multiple vision-language benchmarks

Model Capabilities

Image Understanding and Description
Visual Question Answering
Document Understanding
Chart Analysis
Video Content Understanding
Multimodal Reasoning

Use Cases

Document Processing
Document QA
Extract information from scanned documents or PDFs and answer questions
Achieves 88.0 points on the DocVQA test set
Visual Question Answering
Image Content QA
Answer complex questions about image content
Achieves 79.1 points on the TextVQA validation set
Educational Assistance
Chart Understanding
Interpret and analyze various chart data
Achieves 82.0 points on the ChartQA test set
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
Š 2025AIbase