#!/bin/bash

script to fully prepare ImageNet dataset

1. Download the data

get ILSVRC2012_img_val.tar (about 6.3 GB). MD5: 29b22e2961454d5413ddabcf34fc5622

wget https://image-net.org/data/ILSVRC/2012/ILSVRC2012_img_val.tar

get ILSVRC2012_img_train.tar (about 138 GB). MD5: 1d675b47d978889d74fa0da5fadfb00e

wget https://image-net.org/data/ILSVRC/2012/ILSVRC2012_img_train.tar

2. Extract the training data:

mkdir train && mv ILSVRC2012_img_train.tar train/ && cd train tar -xvf ILSVRC2012_img_train.tar && rm -f ILSVRC2012_img_train.tar find . -name "*.tar" | while read NAME ; do mkdir -p "{NAME%.tar}"; tar -xvf "{NAME}" -C "{NAME%.tar}"; rm -f "{NAME}"; done cd ..

3. Extract the validation data and move images to subfolders:

mkdir val && mv ILSVRC2012_img_val.tar val/ && cd val && tar -xvf ILSVRC2012_img_val.tar wget -qO- https://raw.githubusercontent.com/soumith/imagenetloader.torch/master/valprep.sh | bash

4. Delete corrupted image

there is one png under JPEG name. some readers fail on this image so need to remove it

this line is commented by default to avoid such unexpected behaviour

rm train/n04266014/n04266014_10835.JPEG

5. Sometimes TFRecords may be usefull

wget https://raw.githubusercontent.com/tensorflow/models/master/research/slim/datasets/imagenet_lsvrc_2015_synsets.txt