Adding Vulkan support for executorch on IG4A (pytorch#13806)

Summary:

# Context
Instagram Stories Inspiration and Instagram Computer Vision teams are partnering on bringing  [Vulkan](https://developer.android.com/ndk/guides/graphics/getting-started) support to existing U2Net and squeezeSAM models currently being used to infer saliency and segmentation from images
# This Diff
Adds the necessary dependencies to IG4A to use Vulkan

Reviewed By: hazirbas

Differential Revision: D78291986

Author: rodrigos01meta

extension/android/BUCK

@@ -22,6 +22,7 @@ non_fbcode_target(_kind = fb_android_library,
     language = "JAVA",
     deps = [
         "//fbandroid/java/com/facebook/jni:jni",
+        "//xplat/executorch/backends/vulkan:vulkan_backend_lib_static",
         "//fbandroid/libraries/soloader/java/com/facebook/soloader/nativeloader:nativeloader",
     ],
 )

extension/android/jni/BUCK

@@ -66,6 +66,7 @@ non_fbcode_target(_kind = fb_android_cxx_library,
         "//xplat/executorch/extension/runner_util:inputs_static",
         "//xplat/executorch/extension/tensor:tensor_static",
         "//xplat/executorch/kernels/quantized:generated_lib_static",
+        "//xplat/executorch/backends/vulkan:vulkan_backend_lib_static",
     ],
 )