Alex Williams

Would it be tenable to support both indexing schemes and give users the option to choose? On Tue, Jul 3, 2018, 1:06 AM Hameer Abbasi wrote: > I was thinking...

Yes but if there are performance tradeoffs then it might be nice to support both. Though maybe those tradeoffs are negligible?

The following indexing pattern no longer works on master. I believe the problem originates at commit 222fb8c9ac0ed8cf73dbc6ac2d20dba3f2786fc7: ```python import sparse import numpy as np s = sparse.random((100, 100, 100), density=.1)...

I'd definitely be happy to help with this! (As long as you don't think it would take too much effort on your part to get me started on it.)

Thanks a ton! This is very helpful. How can we fix this? Would it be possible to add something like the following to `Base`: ``` julia convert(::Type{Array{T,2}}, v::Array{Array{T,1},1}) = hcat(v...)'...

This is above my paygrade, but fwiw I think the majority of users would appreciate having a `Matrix` as the default output, especially those (like myself) who are used to...

No I'm after something slightly different. ``` julia using LossFunctions x = LPDistLoss(2) # same as L2DistLoss y = LPDistLoss(1) # same as L1DistLoss # multiple dispatch works for different...

Yes I think importing HDMF accounts for a good chunk of this. Thanks for looking into it.

Thanks - I was hoping for a more automatic solution, but I think you may be right that this is a limitation of matplotlib. On Jun 27, 2016 12:08 PM,...

Would that make softmax `Multivariate{LogitMarginLoss}`? Implementing that seems like a reasonable place to start. If I'm not mistaken it would look a bit like this? ``` julia function value{T