Update on "[ET-VK] High dim tensor support for view, unsqueeze, squeeze, clone"

Differential Revision: [D80800084](https://our.internmc.facebook.com/intern/diff/D80800084)

[ghstack-poisoned]

Author: ssjia

backends/vulkan/runtime/api/containers/Tensor.h

@@ -705,7 +705,8 @@ std::vector<T> flip_and_unsqueeze(
     const size_t numel,
     const int32_t fixed_ndim = -1) {
   const size_t ndim = tensor_metadata.size();
-  size_t ndim_up4 = utils::align_up_4(tensor_metadata.size());
+  size_t ndim_up4 =
+      std::max(utils::align_up_4(tensor_metadata.size()), size_t(4));
 
   if (fixed_ndim > 0) {
     VK_CHECK_COND(fixed_ndim >= ndim);

backends/vulkan/runtime/graph/ops/glsl/view_buffer.glsl

@@ -8,9 +8,6 @@ ${define_required_extensions(DTYPE)}
 
 layout(std430) buffer;
 
-#extension GL_EXT_debug_printf : enable
-
-#define DEBUG_MODE
 #include "indexing.glslh"
 
 ${layout_declare_tensor(B, "w", "t_outp", DTYPE, STORAGE)}
@@ -31,12 +28,6 @@ void main() {
     return;
   }
 
-  bool should_print = outp_bufi == 0;
-  if (should_print) {
-    printBufferMetadata(outp);
-    printBufferMetadata(inp);
-  }
-
   TensorIndex outp_tidx;
   linear_idx_to_tensor_idx(outp, outp_bufi, outp_tidx);
 

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

@@ -18,8 +18,6 @@
 #include <executorch/backends/vulkan/runtime/graph/ops/impl/utils/DimUtils.h>
 #include <executorch/backends/vulkan/runtime/graph/ops/impl/utils/TensorUtils.h>
 
-#include <iostream>
-
 namespace vkcompute {
 
 void add_staging_to_tensor_node(