Implement prototype nodes for each input / output combination #23
Description
There are different possible input / output mappings for Ops, all of which we need to cover with prototype nodes.
- 0-to-1, for
Commands that create new data (eg, a Kernel), see: Support "0-to-n" Commands #24 - 1-to-1, for
Commands that create one output row for one input row (e.g. filter application), see: Support "1 to 1" Commands #25 - 1-to-n, for Ops that create several output rows for one input row (e.g. feature extraction), see: Support "1-to-n" Commands #26