Update base for Update on "[ET-VK][ez] Fix Vulkan Validation layer errors due to consecutive command buffer encoding"

## Changes

* In `VulkanBackend.cpp` do not call `encode_execute()` during model load if the model compile spec specifies `requires_dynamic_shapes` as true
* In test files, do not call `encode_execute()` if `propagate_resize()` is subsequently called.

## Motivation

Recently, it was discovered that a command buffer re-encode was required to update push constant values. This means that for dynamic shapes to work correctly, `encode_execute()` must be called after updating tensor sizes.

As a result, `propagate_resize()` now calls `encode_execute()` internally. This results in scenarios where `encode_execute()` is called once during model load, then again right before the first inference during `propagate_resize()`, without actually executing the command buffer in-between.

This causes Validation layer errors like

```

UNASSIGNED-CoreValidation-DrawState-InvalidImageLayout(ERROR / SPEC): msgNum: 1303270965 - Validation Error: [ UNASSIGNED-CoreValidation-DrawState-InvalidImageLayout ] Object 0: handle = 0x24086224ec0, type = VK_OBJECT_TYPE_COMMAND_BUFFER; Object 1: handle = 0x88d2b500000000e2, type = VK_OBJECT_TYPE_IMAGE; | MessageID = 0x4dae5635 | vkQueueSubmit(): pSubmits[0].pCommandBuffers[0] command buffer VkCommandBuffer 0x24086224ec0[] expects VkImage 0x88d2b500000000e2[] (subresource: aspectMask VK_IMAGE_ASPECT_COLOR_BIT array layer 0, mip level 0) to be in layout VK_IMAGE_LAYOUT_SHADER_READ_ONLY_OPTIMAL--instead, current layout is VK_IMAGE_LAYOUT_UNDEFINED.
    Objects: 2
        [0] 0x24086224ec0, type: 6, name: NULL
        [1] 0x88d2b500000000e2, type: 10, name: NULL
UNASSIGNED-CoreValidation-DrawState-InvalidImageLayout(ERROR / SPEC): msgNum: 1303270965 - Validation Error: [ UNASSIGNED-CoreValidation-DrawState-InvalidImageLayout ] Object 0: handle = 0x24086224ec0, type = VK_OBJECT_TYPE_COMMAND_BUFFER; Object 1: handle = 0x6caffc00000000e3, type = VK_OBJECT_TYPE_IMAGE; | MessageID = 0x4dae5635 | vkQueueSubmit(): pSubmits[0].pCommandBuffers[0] command buffer VkCommandBuffer 0x24086224ec0[] expects VkImage 0x6caffc00000000e3[] (subresource: aspectMask VK_IMAGE_ASPECT_COLOR_BIT array layer 0, mip level 0) to be in layout VK_IMAGE_LAYOUT_SHADER_READ_ONLY_OPTIMAL--instead, current layout is VK_IMAGE_LAYOUT_UNDEFINED.
    Objects: 2
        [0] 0x24086224ec0, type: 6, name: NULL
        [1] 0x6caffc00000000e3, type: 10, name: NULL

```

because the last access information of image/buffer resources are inaccurate during the second command buffer encoding, since the first command buffer never executed.

## Perf Impact

* Performance improvement for first inference of dynamic shape models if actual tensor sizes are much smaller than maximum possible sizes
* No impact for non-dynamic shape models

Differential Revision: [D76047203](https://our.internmc.facebook.com/intern/diff/D76047203/)

cc manuelcandales cbilgin

[ghstack-poisoned]

Author: SS-JIA

.ci/scripts/build_llama_android.sh

@@ -42,7 +42,6 @@ build_llama_runner() {
     popd
     ANDROID_ABI=arm64-v8a
     cmake -DBUCK2="${BUCK2}" \
-    -DBUILD_TESTING=OFF \
     -DCMAKE_TOOLCHAIN_FILE="$ANDROID_NDK"/build/cmake/android.toolchain.cmake  \
     -DANDROID_ABI="${ANDROID_ABI}" \
     -DCMAKE_INSTALL_PREFIX=cmake-android-out \

.ci/scripts/test_llama.sh

@@ -169,7 +169,6 @@ cmake_build_llama_runner() {
     popd
     dir="examples/models/llama"
     retry cmake \
-        -DBUILD_TESTING=OFF \
         -DCMAKE_INSTALL_PREFIX=cmake-out \
         -DCMAKE_BUILD_TYPE="$CMAKE_BUILD_TYPE" \
         -Bcmake-out/${dir} \

.ci/scripts/test_llama_torchao_lowbit.sh

@@ -40,7 +40,6 @@ cmake --build cmake-out -j16 --target install --config Release
 
 # Install llama runner with torchao
 cmake -DPYTHON_EXECUTABLE=python \
-    -DBUILD_TESTING=OFF \
     -DCMAKE_BUILD_TYPE=Release \
     -DEXECUTORCH_BUILD_KERNELS_CUSTOM=ON \
     -DEXECUTORCH_BUILD_KERNELS_OPTIMIZED=ON \

.ci/scripts/test_llava.sh

@@ -64,10 +64,9 @@ cmake_install_executorch_libraries_for_android() {
 
 
 LLAVA_COMMON_CMAKE_ARGS="                        \
-        -DBUILD_TESTING=OFF                      \
         -DPYTHON_EXECUTABLE="$PYTHON_EXECUTABLE" \
         -DCMAKE_INSTALL_PREFIX=${BUILD_DIR}      \
-        -DCMAKE_BUILD_TYPE=${CMAKE_BUILD_TYPE}   \
+        -DCMAKE_BUILD_TYPE=${CMAKE_BUILD_TYPE}         \
         -DEXECUTORCH_BUILD_KERNELS_CUSTOM=ON     \
         -DEXECUTORCH_BUILD_KERNELS_OPTIMIZED=ON  \
         -DEXECUTORCH_BUILD_XNNPACK=ON"

backends/vulkan/test/vulkan_compute_api_test.cpp

@@ -3069,8 +3069,10 @@ void test_to_copy() {
 
   EXPECT_EQ(data_in.size(), output_data.size());
 
+#ifdef VULKAN_DEBUG
   float mse_ex = 0.0f;
   float mse_vk = 0.0f;
+#endif
 
   // check results
   for (size_t i = 0; i < output_data.size(); ++i) {
@@ -3092,6 +3094,9 @@ void test_to_copy() {
         std::bitset<16>(*output_bits).to_string() + ")";
 
     std::cout << msg << std::endl;
+
+    mse_ex += std::pow(expected_output - input, 2);
+    mse_vk += std::pow(output - input, 2);
 #endif
 
     // Note: Torch executor half "rounds up" when converting to fp16 whereas
@@ -3113,13 +3118,12 @@ void test_to_copy() {
     EXPECT_TRUE(
         (*output_bits == *expected_bits) ||
         /*rounding error*/ ((*output_bits + 1u) == *expected_bits));
-    mse_ex += std::pow(expected_output - input, 2);
-    mse_vk += std::pow(output - input, 2);
   }
 
+#ifdef VULKAN_DEBUG
   mse_ex /= output_data.size();
   mse_vk /= output_data.size();
-#ifdef VULKAN_DEBUG
+
   std::cout << "========================================================="
             << std::endl;
   std::cout << "mse_ex = " << mse_ex << ", mse_vk = " << mse_vk << std::endl;

examples/demo-apps/apple_ios/LLaMA/LLaMARunner/LLaMARunner/Exported/LLaMARunner.mm

@@ -14,7 +14,6 @@
 
 using executorch::extension::llm::GenerationConfig;
 using executorch::extension::llm::Image;
-using executorch::extension::llm::TextLLMRunner;
 using executorch::runtime::Error;
 
 NSErrorDomain const LLaMARunnerErrorDomain = @"LLaMARunnerErrorDomain";
@@ -24,15 +23,15 @@ @interface LLaMARunner ()<ExecuTorchLogSink>
 @end
 
 @implementation LLaMARunner {
-  std::unique_ptr<TextLLMRunner> _runner;
+  std::unique_ptr<example::Runner> _runner;
 }
 
 - (instancetype)initWithModelPath:(NSString*)modelPath
                     tokenizerPath:(NSString*)tokenizerPath {
   self = [super init];
   if (self) {
     [ExecuTorchLog.sharedLog addSink:self];
-    _runner = example::create_llama_runner(
+    _runner = example::Runner::create(
         modelPath.UTF8String, tokenizerPath.UTF8String);
   }
   return self;

examples/models/llama/CMakeLists.txt

@@ -220,6 +220,7 @@ endif()
 target_include_directories(
   llama_main
   PUBLIC ${_common_include_directories}
+         ${EXECUTORCH_ROOT}/extension/llm/tokenizers/include
 )
 target_link_libraries(llama_main PUBLIC llama_runner ${link_libraries})
 target_compile_options(llama_main PUBLIC ${_common_compile_options})

examples/models/llama/main.cpp

@@ -81,13 +81,8 @@ int32_t main(int32_t argc, char** argv) {
   }
 #endif
   // create llama runner
-  std::unique_ptr<::executorch::extension::llm::TextLLMRunner> runner =
-      example::create_llama_runner(model_path, tokenizer_path, data_path);
-
-  if (runner == nullptr) {
-    ET_LOG(Error, "Failed to create llama runner");
-    return 1;
-  }
+  std::unique_ptr<example::Runner> runner =
+      example::Runner::create(model_path, tokenizer_path, data_path);
 
   if (warmup) {
     runner->warmup(prompt, /*max_new_tokens=*/seq_len);

examples/models/llama/runner/CMakeLists.txt

@@ -52,20 +52,23 @@ else()
   add_library(llama_runner SHARED ${_llama_runner__srcs})
 endif()
 
-# For extension_llm_runner
-if(NOT TARGET extension_llm_runner)
-  add_subdirectory(
-    ${EXECUTORCH_ROOT}/extension/llm/runner
-    ${CMAKE_CURRENT_BINARY_DIR}/../../../../extension/llm/runner
-  )
-endif()
-
 set(llama_runner_deps executorch_core extension_data_loader extension_module
-                      extension_tensor extension_flat_tensor extension_llm_runner
+                      extension_tensor extension_flat_tensor
 )
 
 target_link_libraries(llama_runner PUBLIC ${llama_runner_deps})
 
+target_include_directories(
+  llama_runner
+  INTERFACE ${_common_include_directories}
+)
+
+# Include tokenizers dependency
+set(CMAKE_POSITION_INDEPENDENT_CODE ON)
+add_subdirectory(
+  ${EXECUTORCH_ROOT}/extension/llm/tokenizers
+  ${CMAKE_CURRENT_BINARY_DIR}/tokenizers
+)
 target_link_libraries(
   llama_runner PUBLIC tokenizers
 )