Add slice-scatter. (#927)

* Add slice-scatter.

* Add the op.

* Make transpose be a no-op when the dimensions are identical.

* Add the backprop.

* And add some gradient test.

candle-core/src/op.rs

@@ -133,6 +133,7 @@ pub enum Op {
     Copy(Tensor),
     Broadcast(Tensor),
     Narrow(Tensor, usize, usize, usize),
+    SliceScatter0(Tensor, Tensor, usize),
     Reshape(Tensor),
     ToDevice(Tensor),
     Transpose(Tensor, usize, usize),