python path fix

Author: winskuo-quic

.github/workflows/pull.yml

@@ -470,10 +470,10 @@ jobs:
 
         # Retrieve 110M Stories Llama Artifacts
         PYTHON_EXECUTABLE=python . .ci/scripts/utils.sh && download_stories_model_artifacts
-        PYTHON_EXECUTABLE -m extension.llm.tokenizer.tokenizer -t tokenizer.model -o tokenizer.bin
+        PYTHONPATH="${PWD}" python -m extension.llm.tokenizer.tokenizer -t tokenizer.model -o tokenizer.bin
         
         # Test static llama stories110m pte size
-        PYTHON_EXECUTABLE=python backends/qualcomm/tests/test_qnn_delegate.py -k TestExampleScript.test_stories_single_llama --model SM8650 --build_folder build-android/ --executorch_root . --artifact_dir . --compile_only"
+        PYTHONPATH="${PWD}" python backends/qualcomm/tests/test_qnn_delegate.py -k TestExampleScript.test_stories_single_llama --model SM8650 --build_folder build-android/ --executorch_root . --artifact_dir . --compile_only"
 
   # Checks accuracy with weight sharing disabled since x86 does not support weight sharing.
   test-static-llama-accuracy-qnn-linux:
@@ -502,10 +502,10 @@ jobs:
 
         # Retrieve 110M Stories Llama Artifacts
         PYTHON_EXECUTABLE=python . .ci/scripts/utils.sh && download_stories_model_artifacts
-        PYTHON_EXECUTABLE -m extension.llm.tokenizer.tokenizer -t tokenizer.model -o tokenizer.bin
+        PYTHONPATH="${PWD}" python -m extension.llm.tokenizer.tokenizer.py -t tokenizer.model -o tokenizer.bin
         
         # Test static llama stories110m accuracy
-        PYTHON_EXECUTABLE=python backends/qualcomm/tests/test_qnn_delegate.py -k TestExampleScript.test_stories_single_llama --model SM8650 --build_folder build-x86_64/ --executorch_root . --artifact_dir . --enable_x86_64"
+        PYTHONPATH="${PWD}" python backends/qualcomm/tests/test_qnn_delegate.py -k TestExampleScript.test_stories_single_llama --model SM8650 --build_folder build-x86_64/ --executorch_root . --artifact_dir . --enable_x86_64"
 
   test-qnn-models-linux:
     name: test-qnn-models-linux