Merge branch 'jz/tt-llama-rebased' into jz/tt-llama-2

Author: jackzhxng

.ci/scripts/test_eval_llama_mmlu.sh

@@ -35,7 +35,7 @@ run_and_verify() {
         exit 1
     fi
     $PYTHON_EXECUTABLE -m examples.models.llama.eval_llama \
-	--model llama2
+	--model llama2 \
 	-c stories110M.pt \
 	-p params.json \
 	-t tokenizer.model \

.ci/scripts/test_llama_runner_eager.sh

@@ -35,7 +35,7 @@ run_and_verify() {
         exit 1
     fi
     $PYTHON_EXECUTABLE -m examples.models.llama.runner.eager \
-	--model llama2
+	--model llama2 \
 	-c stories110M.pt \
 	-p params.json \
 	-t tokenizer.model \

examples/models/llama/export_llama_lib.py

@@ -806,7 +806,7 @@ def _load_llama_model(
     if modelname in EXECUTORCH_DEFINED_MODELS:
         # Set to llama2 because all models in EXECUTORCH_DEFINED_MODELS share the same archteciture as
         # defined in example/models/llama2.
-        modelname = "llama2"
+        modelname = "llama"
         model_class_name = "Llama2Model"
     elif modelname in TORCHTUNE_DEFINED_MODELS:
         if modelname == "llama3_2_vision":