Chiyuan Zhang
Chiyuan Zhang
Can you try to see if calling this function registry_reset(backend::Backend) instead of shutdown and init it would help?
@nikolaypavlov I looked at your code snippet again. Did you call `destroy` on the net after you finish training? If you did not, then the resources will not get released...
Thanks! I'm a bit busy recently. Will take a look at the blas issue when I have a chance. Could you open an issue for that for tracking?
The softmaxloss layer in Mocha is essentially a combination of softmax layer with MultinomialLogisticLoss layer. So if you are to use MultinomialLogisticLossLayer, then SoftmaxLoss layer can be used directly, with...
I fixed some compatibility issue in the master. Could you guys see if it fixed your issue?
What version of Julia and what OS are you using? I tested on Ubuntu and Julia v0.5 release, everything works fine for me.
@mcreel Can you test a stable release (I'm using v0.5.0) instead of testing version like 0.5.1-pre+2?
The line of code reporting `InexactError` is this line: https://github.com/pluskid/Mocha.jl/blob/master/src/layers/pooling/julia-impl.jl#L34 It is trying to assign a value to the mask, which is unsigned. If you try to assign an invalid...
It seems like some pooling region is empty. Just as a sanity check, can you change the kernel for the pooling layer from `(2,1)` to larger values like `(3,1)` to...
@cinvro That is due to previously saved snapshots. Can you remove the saved snapshot files and re-try again? Thanks!