Update on "[ET-VK][Ops] aten.avg_pool2d"

## The Operator
`nn.Module` invocations of [`torch.nn.AvgPool2d`](https://pytorch.org/docs/stable/generated/torch.nn.AvgPool2d.html) get compiled to `aten.avg_pool2d.default` in the Edge Dialect, which carries the following signature.
```
- func: avg_pool2d(Tensor self, int[2] kernel_size, int[2] stride=[], int[2] padding=0, bool ceil_mode=False, bool count_include_pad=True, int? divisor_override=None) -> Tensor
```

## Implementation
This is a full implementation. We start with [LiteInterpreter's `avg_pool2d.glsl` logic](https://github.com/pytorch/pytorch/blob/9257a0698b57acc5607ee6fe31a16fdd93af1731/aten/src/ATen/native/vulkan/glsl/avg_pool2d.glsl), which is incomplete, and cover `ceil_mode=True`,  `count_include_pad=True`, and `divisor_override` cases for full support. As a result, the divisor's computation is now a bit complex. If needed, we can simplify it into separate shaders in the future.

Differential Revision: [D57918523](https://our.internmc.facebook.com/intern/diff/D57918523/)

[ghstack-poisoned]

Author: jorgep31415

backends/vulkan/runtime/graph/ops/impl/utils/KernelUtils.cpp

@@ -156,10 +156,10 @@ std::vector<int64_t> calc_out_sizes_hw(
       make_ivec2_kernel_size(graph, weight, kernel_size_only);
   const auto stride = make_ivec2_from_list(graph, args[0]);
   const auto padding = make_ivec2_from_list(graph, args[1]);
-
   const auto dilation = args[2] == kDummyValueRef
       ? api::utils::ivec2{1, 1}
       : make_ivec2_from_list(graph, args[2]);
+
   if (transposed) {
     const auto output_padding = make_ivec2_from_list(graph, args[3]);
     return calc_transpose_out_sizes_hw(

backends/vulkan/test/test_vulkan_delegate.py

@@ -78,10 +78,6 @@ def assert_outputs_equal(
                     )
         else:
             # If one output, eager returns tensor while executor tuple of size 1
-            # print("ET-VK")
-            # print(model_output[0])
-            # print("Eager-mode")
-            # print(ref_output)
             self.assertTrue(
                 torch.allclose(
                     model_output[0],