Check list for more operators

[mli]

Lacking of operators is the major disadvantage of MXNet comparing to other frameworks. The current multidimensional array interface follows the numpy.ndarray convention. See this tutorial for a quick overview.

Now we are going to add the rest numpy operators into MXNet, and also fix the existing operators if their are different to the according ones in numpy.

I grabbed numpy's routines from this page. The comparison to MXNet's operators will be listed in a few following issues.

For an operator, there are 4 states

• v : already done, and it is consistent to numpy
• p : partially done. the part should be fixed is on the comments
• x : not exists, need to add into mxnet
• = : not exists, but will not support in a short time
• ? : not sure whether x or =, leave for discussion
  We will use separate issues to track each operator category. The schedule is
• survey the current status, and which operators should be added or fixed.
• examples codes to implement new operators
• assign jobs to contributors

Also some related tasks:

• better operator documents
• better unittest
• improve neural network related operators

[tqchen]

close on stale issue