ShixiangWang · September 23, 2021 06:17
diff --git a/clean.sh b/clean.sh
 #!/usr/bin/env bash
 # identify and remove broke images
 # with imageMagick

 for i in $(ls -d train/cat/* train/fish/* val/cat/* val/fish/* test/cat/* test/fish/*); do
    identify $i > /dev/null 2>&1
    if [ $? -ne 0 ]; then
      echo "file $i is broken, removing it"
      rm $i
    fi
 done
diff --git a/download.sh b/download.sh
 #!/usr/bin/env bash
 # https://github.com/Dreamacro/clash/issues/592
 export ALL_PROXY=socks5://127.0.0.1:7891

 #loon batch -f images.csv 'curl -k -o {2}/{1}/"$(basename {0})" -C - {0}'
 loon batch -f images.csv 'out={2}/{1}/"$(basename {0})"; echo processing $out; if ! [ -f $out ]; then curl -k -o $out -C - {0}; fi'
diff --git a/images.csv b/images.csv
	#!/usr/bin/env bash
	# identify and remove broke images
	# with imageMagick

	for i in $(ls -d train/cat/* train/fish/* val/cat/* val/fish/* test/cat/* test/fish/*); do
	identify $i > /dev/null 2>&1
	if [ $? -ne 0 ]; then
	echo "file $i is broken, removing it"
	rm $i
	fi
	done
	#!/usr/bin/env bash
	# https://github.com/Dreamacro/clash/issues/592
	export ALL_PROXY=socks5://127.0.0.1:7891

	#loon batch -f images.csv 'curl -k -o {2}/{1}/"$(basename {0})" -C - {0}'
	loon batch -f images.csv 'out={2}/{1}/"$(basename {0})"; echo processing $out; if ! [ -f $out ]; then curl -k -o $out -C - {0}; fi'
url	class	type
https://farm2.static.flickr.com/1245/1259825348_6a2aa94e8d.jpg	cat	train
https://farm2.static.flickr.com/1080/1029412358_7ee17550fc.jpg	cat	train
https://farm1.static.flickr.com/196/443645811_8c4bb1af50.jpg	cat	train
https://farm2.static.flickr.com/1201/1285591549_593ca7cf6a.jpg	cat	train
https://farm3.static.flickr.com/2002/1760479467_068432dd3f.jpg	cat	train
https://farm2.static.flickr.com/1411/649938210_e9dcbde5ea.jpg	cat	train
https://farm2.static.flickr.com/1423/918775832_9ecfd414b8.jpg	cat	train