gulp>> pytorch-Tiny-Darknet>> 返回
项目作者： Rapternmn

项目描述 ：
PyTorch implementation of the Tiny Darknet Image Classification algorithm
高级语言： Python
项目地址: git://github.com/Rapternmn/pytorch-Tiny-Darknet.git
创建时间： 2019-06-06T11:44:21Z
项目社区：https://github.com/Rapternmn/pytorch-Tiny-Darknet
开源协议：
下载


pytorch-Tiny-Darknet

PyTorch implementation of the Tiny Darknet Image Classification algorithm

This repository contains code for a classifier based on Tiny Darknet, implementedin PyTorch. The code is based on the official code of YOLO v3, as well as a PyTorch
port of the original code, by ayooshkathuria. The aim of this project is to have pytorch compatible Tiny Darknet classifier.

As of now, the code only contains the Classification module.

Requirements

  1. Python 3.5
  2. OpenCV
  3. PyTorch 0.4

Running the detector

On single or multiple images

  1. python detect.py --image imgs/Malamute.jpg

Output would be expected to be something similar like 

  1. Class = malamute  Confidence = 62.50366973876953
  2. Class = Norwegian elkhound  Confidence = 17.33596420288086
  3. Class = Eskimo dog  Confidence = 9.506247520446777
  4. Class = Siberian husky  Confidence = 5.180300235748291
  5. Class = keeshond  Confidence = 4.281615257263184

Inference Time Benchmarks

The average inference time observed on my machine (GTX 1050) is ~ 3ms