From 5d7d54e4d4ce7ab4dc594977fd4357f44394950d Mon Sep 17 00:00:00 2001
From: Jack Zhang <32371937+jackzhxng@users.noreply.github.com>
Date: Wed, 5 Mar 2025 12:23:21 -0800
Subject: [PATCH 1/2] Fix phi4mini test model

---
 .ci/scripts/gather_test_models.py         |  2 +-
 .ci/scripts/test_model.sh                 |  2 +-
 examples/models/__init__.py               |  2 +-
 examples/models/llama/export_llama_lib.py |  2 +-
 examples/models/phi-4-mini/__init__.py    | 14 ++++++++++++++
 5 files changed, 18 insertions(+), 4 deletions(-)
 create mode 100644 examples/models/phi-4-mini/__init__.py

diff --git a/.ci/scripts/gather_test_models.py b/.ci/scripts/gather_test_models.py
index a0cc9efd521..632a185e3b1 100755
--- a/.ci/scripts/gather_test_models.py
+++ b/.ci/scripts/gather_test_models.py
@@ -90,7 +90,7 @@ def model_should_run_on_event(model: str, event: str) -> bool:
     We put higher priority and fast models to pull request and rest to push.
     """
     if event == "pull_request":
-        return model in ["mv3", "vit"]
+        return model in ["mv3", "vit", "phi-4-mini"]
     elif event == "push":
         # These are super slow. Only run it periodically
         return model not in ["dl3", "edsr", "emformer_predict"]
diff --git a/.ci/scripts/test_model.sh b/.ci/scripts/test_model.sh
index 65d04dc3886..c4cb0a8595f 100755
--- a/.ci/scripts/test_model.sh
+++ b/.ci/scripts/test_model.sh
@@ -100,7 +100,7 @@ test_model() {
       rm "./${MODEL_NAME}.pte"
       return  # Skip running with portable executor runnner since portable doesn't support Qwen's biased linears.
   fi
-  if [[ "${MODEL_NAME}" == "phi4_mini" ]]; then
+  if [[ "${MODEL_NAME}" == "phi-4-mini" ]]; then
       # Install requirements for export_llama
       bash examples/models/llama/install_requirements.sh
       # Test export_llama script: python3 -m examples.models.llama.export_llama.
diff --git a/examples/models/__init__.py b/examples/models/__init__.py
index 778138a1802..26c84c48fc6 100644
--- a/examples/models/__init__.py
+++ b/examples/models/__init__.py
@@ -35,7 +35,7 @@
     "llava": ("llava", "LlavaModel"),
     "efficient_sam": ("efficient_sam", "EfficientSAM"),
     "qwen2_5": ("qwen2_5", "Qwen2_5Model"),
-    "phi4_mini": ("phi4_mini", "Phi4MiniModel"),
+    "phi-4-mini": ("phi-4-mini", "Phi4MiniModel"),
 }
 
 __all__ = [
diff --git a/examples/models/llama/export_llama_lib.py b/examples/models/llama/export_llama_lib.py
index 6e5eca11d0e..2319ec0c6a7 100644
--- a/examples/models/llama/export_llama_lib.py
+++ b/examples/models/llama/export_llama_lib.py
@@ -93,7 +93,7 @@
     "llama3_2",
     "static_llama",
     "qwen2_5",
-    "phi4_mini",
+    "phi-4-mini",
 ]
 TORCHTUNE_DEFINED_MODELS = ["llama3_2_vision"]
 
diff --git a/examples/models/phi-4-mini/__init__.py b/examples/models/phi-4-mini/__init__.py
new file mode 100644
index 00000000000..056f2c26314
--- /dev/null
+++ b/examples/models/phi-4-mini/__init__.py
@@ -0,0 +1,14 @@
+# This source code is licensed under the BSD-style license found in the
+# LICENSE file in the root directory of this source tree.
+
+from executorch.examples.models.llama.model import Llama2Model
+
+
+class Phi4MiniModel(Llama2Model):
+    def __init__(self, **kwargs):
+        super().__init__(**kwargs)
+
+
+__all__ = [
+    "Phi4MiniModel",
+]

From 87a78df85e5b420db8e603e6757f755f59f6645f Mon Sep 17 00:00:00 2001
From: Jack Zhang <32371937+jackzhxng@users.noreply.github.com>
Date: Wed, 5 Mar 2025 16:27:35 -0800
Subject: [PATCH 2/2] Remove pull trigger since done testing

---
 .ci/scripts/gather_test_models.py | 2 +-
 .ci/scripts/test_model.sh         | 1 +
 2 files changed, 2 insertions(+), 1 deletion(-)

diff --git a/.ci/scripts/gather_test_models.py b/.ci/scripts/gather_test_models.py
index 632a185e3b1..a0cc9efd521 100755
--- a/.ci/scripts/gather_test_models.py
+++ b/.ci/scripts/gather_test_models.py
@@ -90,7 +90,7 @@ def model_should_run_on_event(model: str, event: str) -> bool:
     We put higher priority and fast models to pull request and rest to push.
     """
     if event == "pull_request":
-        return model in ["mv3", "vit", "phi-4-mini"]
+        return model in ["mv3", "vit"]
     elif event == "push":
         # These are super slow. Only run it periodically
         return model not in ["dl3", "edsr", "emformer_predict"]
diff --git a/.ci/scripts/test_model.sh b/.ci/scripts/test_model.sh
index c4cb0a8595f..5940affc109 100755
--- a/.ci/scripts/test_model.sh
+++ b/.ci/scripts/test_model.sh
@@ -107,6 +107,7 @@ test_model() {
       "${PYTHON_EXECUTABLE}" -m examples.models.llama.export_llama --model "${MODEL_NAME}" -c examples/models/llama/params/demo_rand_params.pth -p examples/models/phi-4-mini/config.json
       run_portable_executor_runner
       rm "./${MODEL_NAME}.pte"
+      return
   fi
 
   # Export a basic .pte and run the model.