Update on "Use generated files in module test"

Instead of checking in the PTE/PTD files. This helps us keep up to date with aot changes, especially with incoming flat_tensor refactor.

Main changes here are with the model. Instead of using AddModule from examples/portable, use ModuleAdd from test/models/export_program, like our other tests. There is a shape difference (1 tensor vs 2 tensors+alpha), so we have to change all the inputs.

Differential Revision: [D73470865](https://our.internmc.facebook.com/intern/diff/D73470865/)

[ghstack-poisoned]

Author: lucylq

extension/module/test/CMakeLists.txt

@@ -29,7 +29,6 @@ add_custom_command(
   COMMAND
     python3 -m test.models.export_program --modules "ModuleLinear"
     --external-constants --outdir "${CMAKE_CURRENT_BINARY_DIR}" 2> /dev/null
-  WORKING_DIRECTORY ${EXECUTORCH_ROOT}
 )
 
 add_custom_target(

extension/module/test/module_test.cpp

@@ -23,10 +23,6 @@ using namespace ::executorch::runtime;
 class ModuleTest : public ::testing::Test {
  protected:
   static void SetUpTestSuite() {
-    std::string resources_path;
-    if (const char* env = std::getenv("RESOURCES_PATH")) {
-      resources_path = env;
-    }
     model_path_ = std::getenv("ET_MODULE_ADD_PATH");
     linear_path_ = std::getenv("ET_MODULE_LINEAR_PROGRAM_PATH");
     linear_data_path_ = std::getenv("ET_MODULE_LINEAR_DATA_PATH");