Kanvi/fused BMMT support

[kanvi-nervana]

NGraph now supports a fused BatchMatMulTranspose. This PR is make changes in the translation for BatchMatMul to accommodate that.

[sayantan-nervana]

LGTM.

General comments:

Since this translate op will be heavily used, lets bump up its error messages:

For example:
if (ng_lhs_shape.size() != ng_rhs_shape.size()) {
return errors::InvalidArgument(
"Dimensions of two input args are not the same for BatchMatMul");
}

could become:

if (ng_lhs_shape.size() != ng_rhs_shape.size()) {
return errors::InvalidArgument(
"Dimensions of two input args are not the same for BatchMatMul. Left shape is ", ng::join(ng_lhs_shape), " of rank ", ng_lhs_shape.size(), " and right shape is ", ng::join(ng_rhs_shape), " of rank ", ng_rhs_shape.size());
}

similarly for ""Dimensions of input args for BatchMatMul must be >=2", n_dims);" we can say : "Dimensions of input args for BatchMatMul must be >=2 but is ", n_dims);

and:
"ng_lhs_shape and ng_rhs_shape must be the same for BatchMatMul "
"for each dimension ",
i)
could be:
"ng_lhs_shape and ng_rhs_shape must be the same for BatchMatMul "
"for each dimension but found ",
i, "th dimension different.", ) //and print teh left shape and rigth shape using ng::join()

[sayantan-nervana]

Maybe we do not need the else parts:

else {
      ng_lhs_axes.push_back(n_dims - 2);
      ng_lhs_axes.push_back(n_dims - 1);
    }

[sayantan-nervana]

I'd say we can construct ng_lhs_axes and ng_rhs_axes in each of the cases using iota instead of push_back (which is slower probably) and get rid of out_axes

for example in case n>3, only 1 needs to be constructed. we do not need both ng_lhs_axes and ng_rhs_axes

[sayantan-nervana]

Computing ng_lhs_shape = ng_lhs->get_shape(); is not needed in n==2 case. There is some leftover/legacy code from the previous implementation that we can clean up

ng_lhs = ng::builder::numpy_transpose(ng_lhs, tf_adj_x ? {1,0} : {0,1}); // or ng::AxisVector{0,1}
Builder::SetTracingInfo(op->name(), ng_lhs);
ng_rhs = ng::builder::numpy_transpose(ng_lhs, tf_adj_y ? {1,0} : {0,1});
Builder::SetTracingInfo(op->name(), ng_rhs);
SaveNgOp(ng_op_map, op->name(),
             ConstructNgNode<ngraph::op::Dot>(op->name(), ng_lhs, ng_rhs));

[kanvi-nervana]

Done

[sayantan-nervana]

these 2 are not needed/used for n_dims == 2
ng_lhs_shape = ng_lhs->get_shape();
ng_rhs_shape = ng_rhs->get_shape();

[sayantan-nervana]

LGTM