Chewxy

_From @kabaka0 on July 3, 2017 23:20_ My reasoning was that there could be cases in which a scalar is masked, such as, say, a reduction operation from a masked...

_From @kabaka0 on July 3, 2017 23:26_ Also in the case you quote, `retVal = New(FromScalar(argmaxI(t.Ints(), t.mask)))`, the return value is a regular (unmasked) scalar, the mask is only used...

OK. Thanks. Very helpful. Can you craft a documentation for this? Thanks

_From @kabaka0 on July 3, 2017 23:35_ Sure, I will do! On Jul 3, 2017 7:34 PM, "Chewxy" wrote: > Assigned #132 to @kabaka0 > . > > — >...

(btw, just send a PR, I'll merge your documentation before #127

_From @kabaka0 on June 28, 2017 21:8_ I can see the logic behind the split of tensor - the package is currently quite large, which made it harder to locate...

I was thinking precisely embedding ``` type MA struct { *Dense mask []bool ... } ```

(I'll do the heavy lifting.. given I'm working on the sparse and colmajor Tensors) you just provide the struct you have in mind @kabaka0

_From @kabaka0 on June 28, 2017 21:15_ That works. ``` type MA struct { *Dense mask []bool fillValue interface{} ... } ```

_From @kabaka0 on June 28, 2017 21:16_ I am happy to help with it, but if you are actively editing the package then I can leave it to you for...