Update on "[ET-VK][Ops] quantize_per_channel reference impl and testing"

# Context

In order to properly enable dynamic quantization, we create the quantize_per_channel operator as its seemingly useful to have for the pipeline.

# Changes

This creates the wrapper for the cpu reference implementation, and also a dummy reference implementation I created just to test against it.

Differential Revision: [D77746132](https://our.internmc.facebook.com/intern/diff/D77746132/)

[ghstack-poisoned]

Author: morelos

backends/vulkan/test/op_tests/quantize_test.cpp

@@ -473,18 +473,6 @@ void test_vulkan_quantize_per_token_impl(
     const vkcompute::utils::StorageType in_storage,
     const vkcompute::utils::StorageType out_storage);
 
-void test_vulkan_quantize_per_channel_impl(
-    const std::vector<int>& input_sizes,
-    const std::vector<float>& scales,
-    const std::vector<int>& zero_points,
-    int64_t axis,
-    int64_t quant_min,
-    int64_t quant_max,
-    at::ScalarType in_dtype,
-    at::ScalarType dtype,
-    const vkcompute::utils::StorageType in_storage,
-    const vkcompute::utils::StorageType out_storage);
-
 // Wrapper function to test both buffer and texture storage types
 void test_vulkan_quantize_per_tensor(
     const std::vector<int>& input_sizes,
@@ -565,48 +553,6 @@ void test_vulkan_quantize_per_token(
       vkcompute::utils::kTexture3D);
 }
 
-// Wrapper function to test both buffer and texture storage types
-void test_vulkan_quantize_per_channel(
-    const std::vector<int>& input_sizes,
-    const std::vector<float>& scales,
-    const std::vector<int>& zero_points,
-    int64_t axis,
-    int64_t quant_min,
-    int64_t quant_max,
-    at::ScalarType in_dtype = at::kFloat,
-    at::ScalarType dtype = at::kInt) {
-  // Test with buffer storage
-  test_vulkan_quantize_per_channel_impl(
-      input_sizes,
-      scales,
-      zero_points,
-      axis,
-      quant_min,
-      quant_max,
-      in_dtype,
-      dtype,
-      vkcompute::utils::kBuffer,
-      vkcompute::utils::kBuffer);
-
-  // If the in_dtype is a double, convert to float for texture implementation
-  // since they don't support 64bit as inputs
-  if (in_dtype == at::kDouble) {
-    in_dtype = at::kFloat;
-  }
-
-  test_vulkan_quantize_per_channel_impl(
-      input_sizes,
-      scales,
-      zero_points,
-      axis,
-      quant_min,
-      quant_max,
-      in_dtype,
-      dtype,
-      vkcompute::utils::kTexture3D,
-      vkcompute::utils::kTexture3D);
-}
-
 void test_reference_quantize_per_tensor(
     const std::vector<int>& input_sizes,
     float scale,