Update on "[ET-VK][12/n] aten.cat with more codegen"

1. The `aten.cat` operation are very straightforward using the `copy_*_node`.
2. Complexity comes from the code-gen. We need to introduce a `AT_TENSOR_LIST` type, which contains a list of `AT_TENSOR` with `is_in=True`. The tensor list itself as a container is not an `IOValueRef`, but the element inside are. It leads to some ugly if-then-else in the codegen.

Differential Revision: [D56626865](https://our.internmc.facebook.com/intern/diff/D56626865/)

[ghstack-poisoned]

Author: yipjustin

backends/vulkan/runtime/graph/ops/impl/Cat.cpp

@@ -33,12 +33,6 @@ void add_cat_default_node(
   int64_t dim = graph.extract_scalar<int64_t>(dim_ref);
   vTensorPtr t_out = graph.get_tensor(out);
 
-  /*
-  for (ValueRef input_ref : *input_list) {
-    vTensorPtr t_in = graph.get_tensor(input_ref);
-  }
-  */
-
   NchwDim nchw_dim = normalize_to_nchw_dim(*t_out, dim);
 
   // TODO: Find ways to factor out the similar code for width, height, and batch