Update on "Reuse GELU implementation from PyTorch core"

kernels/optimized doesn't need to support embedded systems, so it can just take a header-only dep on PyTorch.

Note that, because we will pick up Sleef internally and ignore it
externally thanks to ATen vec, this PR gets to enable optimized GELU in OSS.

Testing: CI to make sure this doesn't break mobile build modes; happy to take advice on anything not currently covered that might break.

Differential Revision: [D66335522](https://our.internmc.facebook.com/intern/diff/D66335522/)

[ghstack-poisoned]

Author: Github Executorch

.ci/scripts/build_llama_android.sh

@@ -10,6 +10,11 @@ set -exu
 # shellcheck source=/dev/null
 source "$(dirname "${BASH_SOURCE[0]}")/utils.sh"
 
+if [[ -z "${PYTHON_EXECUTABLE:-}" ]]; then
+  PYTHON_EXECUTABLE=python3
+fi
+which "${PYTHON_EXECUTABLE}"
+
 install_executorch_and_backend_lib() {
   echo "Installing executorch and xnnpack backend"
   clean_executorch_install_folders

.ci/scripts/utils.sh

@@ -136,7 +136,7 @@ cmake_install_executorch_lib() {
   clean_executorch_install_folders
   retry cmake -DBUCK2="$BUCK" \
           -DCMAKE_INSTALL_PREFIX=cmake-out \
-          -DCMAKE_PREFIX_PATH="$($PYTHON -c 'import torch as _; print(_.__path__[0])')" \
+          -DCMAKE_PREFIX_PATH="$($PYTHON_EXECUTABLE -c 'import torch as _; print(_.__path__[0])')" \
           -DCMAKE_BUILD_TYPE=Release \
           -DPYTHON_EXECUTABLE="$PYTHON_EXECUTABLE" \
           -Bcmake-out .

test/run_oss_cpp_tests.sh

@@ -22,6 +22,11 @@ elif [[ $(uname) == "Linux" ]]; then
   export LLVM_COV="${LLVM_COV:-llvm-cov}"
 fi
 
+if [[ -z "${PYTHON_EXECUTABLE:-}" ]]; then
+  PYTHON_EXECUTABLE=python3
+fi
+which "${PYTHON_EXECUTABLE}"
+
 build_executorch() {
   BUILD_VULKAN="OFF"
   if [ -x "$(command -v glslc)" ]; then