11:27 < tham> Could anyone provide some examples of cnn?I would like to use it to do some image classification tasks

11:33 < tham> About the sparse_autoencoder I commit yesterday, I put it in ann module because this class will break the old codes.Besides, I hope it could leverage different layers of ann

11:34 < tham> But I am not sure my implementation is correct

11:38 < tham> The codes are based on the implementaion of the original SparseAutoencoder, just some minor chnage

11:38 < tham> I will try to write some test codes and upload to the server

11:39 < tham> Have no idea why the pull request would cause failed 249 yet

12:39 < tham> I upload the test case

12:40 < tham> the file name is sparse_autoencoder_test_2.cpp

12:40 < tham> it is based on the original test cases from sparse_autoencoder_test.cpp

12:56 < naywhayare> hi tham, thanks for the tests -- I was going to port the sparse_autoencoder_test.cpp tests but you beat me to it :)

12:56 < naywhayare> I was going to merge it yesterday, but since it's part of the ann code, I'll let zoq do it since he's far more familiar with that code

12:57 < naywhayare> the build is failing because you have two test suites named "SparseAutoencoderTest"

12:57 < naywhayare> i.e. change line 30 of sparse_autoencoder_test_2.cpp

13:23 < tham> Okay, I change the name of it, but the error I mentioned about is 249, the error you mentioned about is 251

13:23 < tham> Do zoq have any plan to write some examples?

13:24 < tham> I heard that cnn is quite success on image and voice recognition

13:24 < tham> It would be nice if there are some examples

13:28 < naywhayare> tham: ah, #249, sorry about the confusion

13:28 < naywhayare> it looks like the build was unlucky. many of the mlpack tests are probabilistic, so sometimes they will randomly fail

13:28 < naywhayare> this usually means that the tolerances need to be loosened a little bit

13:29 < naywhayare> but there is nothing wrong with your code in that instance

13:29 < naywhayare> I've seen that test (NystroemMethodTest/Rank10Test) fail a few times now... I'll see if I can figure out where that NaN is coming from

13:31 < tham> Thanks, I recommit, this time it should be able to compile