Pelee

Visit Tool

Pelee is an open-source object detection system designed for real-time performance on mobile devices. It provides code for implementing and experimenting with the Pelee object detection model.

Claim this tool

1View

At a glance

Pricing

Open Source

Free tier

Yes

API

Skill level

Technical

About

What is Pelee?

Pelee is an open-source, real-time object detection system specifically optimized for mobile devices. Based on the SSD framework, this system was presented at NeurIPS 2018 and offers a robust solution for mobile vision applications. The repository provides the necessary code for researchers and developers to implement and experiment with the Pelee object detection model. It includes tools for training on VOC and COCO datasets, evaluation scripts, and pre-trained models for various configurations. Pelee demonstrates competitive performance in terms of both accuracy (mAP) and frames per second (FPS) on mobile platforms like iPhone 8 and NVIDIA TX2, making it a valuable resource for developing efficient mobile AI applications.

Best used for

Ideal for developers who need to implement efficient real-time object detection on mobile devices, conduct research in mobile computer vision, and benchmark performance on various mobile platforms. Especially valuable for those working with the SSD framework and requiring pre-trained models for VOC and COCO datasets.

Common actions

detect objects

implement computer vision

optimize mobile AI

train object detection models

github copilot"AI Agents"automated workflowface swappingopen-sourceworkflowscollaborationdeepfakelow-code/no-code

Capabilities

Key features

Real-time object detection
Mobile device optimized
SSD framework base
Pre-trained models
VOC/COCO dataset support
Evaluation scripts

Target Audience

developer

Integrations

Not yet documented

Pricing & Plans

Open Source

Free

FAQs

What datasets does Pelee support for training and evaluation?

Pelee supports training and evaluation on both PASCAL VOC 2007/2012 and MS COCO datasets. The repository provides specific scripts and instructions for preparing these datasets and training the Pelee model accordingly.

What are the performance characteristics of Pelee on mobile devices?

Pelee is optimized for real-time performance on mobile devices. Benchmarks show it achieves 23.6 FPS on iPhone 8 and 125 FPS on NVIDIA TX2, with a mAP of 70.9% on VOC 2007, demonstrating its efficiency.

How can I get started with training a Pelee model?

To train a Pelee model, you need to install SSD Caffe, download the PASCAL VOC datasets, create LMDB files, and then clone the Pelee repository. Training scripts like `train_voc.py` are provided for different datasets.

Trending

Subcategories trending in Coding & Development

Open Source & Models DevOps & Infrastructure No-Code / Low-Code Testing & QA Backend & APIs Prompt Engineering

Trending

Explore

Browse AI tools by category

Content & Design Productivity & Business Coding & Development AI Agents & Automation Research & Education Wellness & Lifestyle Career Development Marketing & Growth Data & Analytics Customer Support & CX Finance E-commerce