NetTrain always crashes in out of core training with NetEncoder Setting

Asked Jun 20 '17 at 03:29

Active Jun 21 '17 at 15:30

Viewed 232 times

See this official example example

My enviroment is Windows 10 GTX1080 with Mathematica 11.1.1

It's Ok if I import all training images into memory.

Even in the simple caseNetTrain[net, trainningData] with only several samples.

Seems the problem is about NetEncoder[{"Image", {100, 35}], change the ImageDimensions will reproduce the error. ImageDimensions in NetEncoder should be the same[35,35]?

OK in trained = NetTrain[net, MapAt[Import@# &, trainingData, {All, 1}]]

Bug?

edited Jun 21 '17 at 15:30

asked Jun 20 '17 at 03:29

HyperGroups

8,619
1
26
63

Could you try this example and see if it works? – xslittlegrass Jun 20 '17 at 15:11
@xslittlegrass OK, in this official example... – HyperGroups Jun 21 '17 at 04:07
@xslittlegrass seems a bug. – HyperGroups Jun 21 '17 at 15:25

NetTrain always crashes in out of core training with NetEncoder Setting

0 Answers0

Linked