Commitaf7dff5

committed

fix issue on yolov3 tutorial

1 parent3105f40 commitaf7dff5Copy full SHA for af7dff5

File tree

+15

-3

lines changed

+15

-3

lines changed

Lines changed: 5 additions & 1 deletion

Original file line number	Diff line number	Diff line change
`@@ -16,7 +16,11 @@`
`16`	`16`	`net=cv2.dnn.readNetFromDarknet(config_path,weights_path)`
`17`	`17`
`18`	`18`	`ln=net.getLayerNames()`
`19`		`-ln= [ln[i[0]-1]foriinnet.getUnconnectedOutLayers()]`
	`19`	`+try:`
	`20`	`+ln= [ln[i[0]-1]foriinnet.getUnconnectedOutLayers()]`
	`21`	`+exceptIndexError:`
	`22`	`+# in case getUnconnectedOutLayers() returns 1D array when CUDA isn't available`
	`23`	`+ln= [ln[i-1]foriinnet.getUnconnectedOutLayers()]`
`20`	`24`
`21`	`25`	`cap=cv2.VideoCapture(0)`
`22`	`26`

Lines changed: 5 additions & 1 deletion

Original file line number	Diff line number	Diff line change
`@@ -17,7 +17,11 @@`
`17`	`17`	`net=cv2.dnn.readNetFromDarknet(config_path,weights_path)`
`18`	`18`
`19`	`19`	`ln=net.getLayerNames()`
`20`		`-ln= [ln[i[0]-1]foriinnet.getUnconnectedOutLayers()]`
	`20`	`+try:`
	`21`	`+ln= [ln[i[0]-1]foriinnet.getUnconnectedOutLayers()]`
	`22`	`+exceptIndexError:`
	`23`	`+# in case getUnconnectedOutLayers() returns 1D array when CUDA isn't available`
	`24`	`+ln= [ln[i-1]foriinnet.getUnconnectedOutLayers()]`
`21`	`25`	`# read the file from the command line`
`22`	`26`	`video_file=sys.argv[1]`
`23`	`27`	`cap=cv2.VideoCapture(video_file)`

Lines changed: 5 additions & 1 deletion

Original file line number	Diff line number	Diff line change
`@@ -37,7 +37,11 @@`
`37`	`37`
`38`	`38`	`# get all the layer names`
`39`	`39`	`ln=net.getLayerNames()`
`40`		`-ln= [ln[i[0]-1]foriinnet.getUnconnectedOutLayers()]`
	`40`	`+try:`
	`41`	`+ln= [ln[i[0]-1]foriinnet.getUnconnectedOutLayers()]`
	`42`	`+exceptIndexError:`
	`43`	`+# in case getUnconnectedOutLayers() returns 1D array when CUDA isn't available`
	`44`	`+ln= [ln[i-1]foriinnet.getUnconnectedOutLayers()]`
`41`	`45`	`# feed forward (inference) and get the network output`
`42`	`46`	`# measure how much it took in seconds`
`43`	`47`	`start=time.perf_counter()`

Comments

(0)