Pin bump for optimum-executorch (#12022)

### Summary
Pin bump to include latest changes for gemma-3, phi-4, and more

### Test plan
CI

---------

Co-authored-by: Guang Yang <[email protected]>

Author: guangy10

.ci/docker/ci_commit_pins/optimum-executorch.txt

@@ -0,0 +1 @@
+a3942627f5ac048e06b4b1d703b0a6a53bf6da5b

.github/workflows/android-perf.yml

@@ -341,10 +341,11 @@ jobs:
               echo "tokenizer.json is downloaded to $DOWNLOADED_PATH"
 
               # Install optimum-executorch
+              OPTIMUM_ET_COMMIT=$(cat .ci/docker/ci_commit_pins/optimum-executorch.txt)
               git clone https://github.com/huggingface/optimum-executorch
               pushd optimum-executorch
               # There is no release yet, for CI stability, always test from the same commit on main
-              git checkout 4c3b18f6cca68c5ccff809131d570062723d7188
+              git checkout $OPTIMUM_ET_COMMIT
               python install_dev.py --skip_override_torch
               pip list
 
@@ -353,21 +354,12 @@ jobs:
                 "--task" "text-generation"
                 "--recipe" "xnnpack"
                 "--use_custom_sdpa"
+                "--use_custom_kv_cache"
                 "--qlinear"
                 "--qembedding"
                 "--output_dir" ".."
               )
 
-              # Add conditional arguments based on model
-              case "${HF_MODEL_REPO}" in
-                *"google/gemma-3-1b-it"*)
-                  echo "--use_custom_kv_cache can not be used for HybridCache"
-                  ;;
-                *)
-                  ARGS+=("--use_custom_kv_cache")
-                  ;;
-              esac
-
               optimum-cli export executorch "${ARGS[@]}"
               popd
 

.github/workflows/apple-perf.yml

@@ -346,10 +346,11 @@ jobs:
             echo "tokenizer.json is downloaded to $DOWNLOADED_PATH"
 
             # Install optimum-executorch
+            OPTIMUM_ET_COMMIT=$(cat .ci/docker/ci_commit_pins/optimum-executorch.txt)
             git clone https://github.com/huggingface/optimum-executorch
             pushd optimum-executorch
             # There is no release yet, for CI stability, always test from the same commit on main
-            git checkout 4c3b18f6cca68c5ccff809131d570062723d7188
+            git checkout $OPTIMUM_ET_COMMIT
             ${CONDA_RUN} python install_dev.py --skip_override_torch
             pip list
 
@@ -358,21 +359,12 @@ jobs:
               "--task" "text-generation"
               "--recipe" "xnnpack"
               "--use_custom_sdpa"
+              "--use_custom_kv_cache"
               "--qlinear"
               "--qembedding"
               "--output_dir" ".."
             )
 
-            # Add conditional arguments based on model
-            case "${HF_MODEL_REPO}" in
-              *"google/gemma-3-1b-it"*)
-                echo "--use_custom_kv_cache can not be used for HybridCache"
-                ;;
-              *)
-                ARGS+=("--use_custom_kv_cache")
-                ;;
-            esac
-
             ${CONDA_RUN} optimum-cli export executorch "${ARGS[@]}"
             popd
 

.github/workflows/trunk.yml

@@ -594,10 +594,11 @@ jobs:
         echo "::group::Set up Hugging Face"
         pip install -U "huggingface_hub[cli]"
         huggingface-cli login --token $SECRET_EXECUTORCH_HF_TOKEN
+        OPTIMUM_ET_COMMIT=$(cat .ci/docker/ci_commit_pins/optimum-executorch.txt)
         git clone https://github.com/huggingface/optimum-executorch
         pushd optimum-executorch
         # There is no release yet, for CI stability, always test from the same commit on main
-        git checkout 4c3b18f6cca68c5ccff809131d570062723d7188
+        git checkout $OPTIMUM_ET_COMMIT
         python install_dev.py --skip_override_torch
         popd
         pip list
@@ -614,21 +615,12 @@ jobs:
           "--task" "text-generation"
           "--recipe" "xnnpack"
           "--use_custom_sdpa"
+          "--use_custom_kv_cache"
           "--qlinear"
           "--qembedding"
           "--output_dir" "${OUTPUT_DIR}"
         )
 
-        # Add conditional arguments based on model
-        case "${MODEL_ID}" in
-          *"google/gemma-3-1b-it"*)
-            echo "--use_custom_kv_cache can not be used for HybridCache"
-            ;;
-          *)
-            ARGS+=("--use_custom_kv_cache")
-            ;;
-        esac
-
         optimum-cli export executorch "${ARGS[@]}"
 
         ls -FlAGhp ${OUTPUT_DIR}
@@ -732,18 +724,18 @@ jobs:
       timeout: 90
       script: |
         set -eux
-        
+
         # The generic Linux job chooses to use base env, not the one setup by the image
         CONDA_ENV=$(conda env list --json | jq -r ".envs | .[-1]")
         conda activate "${CONDA_ENV}"
-        
+
         # Build and install Executorch
         PYTHON_EXECUTABLE=python \
         CMAKE_ARGS="-DEXECUTORCH_BUILD_NXP_NEUTRON=ON" \
         .ci/scripts/setup-linux.sh --build-tool "cmake"
-        
+
         # Install test requirements
         pip install -r backends/nxp/requirements-tests.txt
-        
+
         # Run pytest
         PYTHON_EXECUTABLE=python bash backends/nxp/run_unittests.sh