rscup2019_classifier

rscup2019，classifier challenge, baseline

中文版本

Introduction

RSCUP: 遥感图像场景分类

Accuracy

This baseline got 0.893

Requirements

pytorch==1.1 (1.0 should OK)

tensorboard==1.8

tensorboardX==1.7

pillow

Memory>=8G

GPU=1080Ti

Usage

The simplest way to use it is simply to specify the data_dir parameter. You will get 0.893 after 85m, also you will get log and tensorboard_log.

python train_resnet.py --data_dir your_data_dir

your data dir should be organized as follows:

data_dir/
    |->train
    |->val
    |->test
    |->ClsName2id.txt

tensorboard