rcurtin_irc changed the topic of #mlpack to: mlpack: a scalable machine learning library (https://www.mlpack.org/) -- channel logs: https://libera.irclog.whitequark.org/mlpack -- NOTE: messages sent here might not be seen by bridged users on matrix, gitter, or slack
<jonpsy[m]> <EshaanAgarwal[m]> "i did try subtracting old and..." <- Hey, sorry office work came up & it was getting pretty late.
<jonpsy[m]> <EshaanAgarwal[m]> "But how would that let me know..." <- I mean, you can set a filter that if diff is less than 1E-7, consider it zero
<zoq[m]> The weights are different or the gradient?
<jonpsy[m]> Both, the actor gradients are crazy
<jonpsy[m]> its `[0, 0, 0, ......., +x, -x]`
<jonpsy[m]> this pattern is repeating...
<zoq[m]> But the forward pass is correct?
<zoq[m]> activations?
<jonpsy[m]> Not sure, bcuz he hasn't yet fixed initial state for both `pytorch` & `mlpack`
<zoq[m]> Eshaan Agarwal: what is the issue with fixing the state for both?
<rcurtin[m]> shrit: I handled all the comments you left on the namespaces PR, let me know if you think there is anything else that needs to be done :) and *thanks* for taking the huge amount of time to review that!