diff --git a/.github/ci_expected_accuracy/check_expected.py b/.github/ci_expected_accuracy/check_expected.py
index 48543c930..6d321e34b 100644
--- a/.github/ci_expected_accuracy/check_expected.py
+++ b/.github/ci_expected_accuracy/check_expected.py
@@ -16,14 +16,14 @@
 
 
 # load csv files
-test_data= pd.read_csv(args.csv_file)
+test_data= pd.read_csv(args.csv_file, comment='#')
 # test_data = test_data.reset_index()  # make sure indexes pair with number of rows
 # test_data = test_data.sort_values(by=["name"], ascending=True)
 test_names = [row["name"] for index, row in test_data.iterrows()]
 
 current_path = pathlib.Path(__file__).parent.resolve()
 refer_file = str(current_path) + "/" + args.category + "_" + args.suite + "_" + args.mode + ".csv"
-refer_data= pd.read_csv(refer_file)
+refer_data= pd.read_csv(refer_file, comment='#')
 # refer_data = refer_data.reset_index()  # make sure indexes pair with number of rows
 # refer_data = refer_data.sort_values(by=["name"], ascending=True)
 refer_names = [row["name"] for index, row in refer_data.iterrows()]
diff --git a/.github/ci_expected_accuracy/inductor_torchbench_inference.csv b/.github/ci_expected_accuracy/inductor_torchbench_inference.csv
index 4825aa41f..832923854 100644
--- a/.github/ci_expected_accuracy/inductor_torchbench_inference.csv
+++ b/.github/ci_expected_accuracy/inductor_torchbench_inference.csv
@@ -102,5 +102,6 @@ torch_multimodal_clip,pass,pass,pass,eager_fail_to_run,eager_fail_to_run
 tts_angular,pass,eager_fail_to_run,eager_fail_to_run,eager_fail_to_run,eager_fail_to_run
 vgg16,pass,pass,pass,pass,pass
 vision_maskrcnn,pass,pass,pass,eager_fail_to_run,eager_fail_to_run
-yolov3,pass,pass,pass,pass,pass
+# Skip yolov3 for known torchbench issue: https://github.com/intel/torch-xpu-ops/issues/1229
+yolov3,eager_fail_to_run,eager_fail_to_run,eager_fail_to_run,eager_fail_to_run,eager_fail_to_run
 hf_Roberta_base,pass,pass,pass,pass,pass
diff --git a/.github/ci_expected_accuracy/inductor_torchbench_training.csv b/.github/ci_expected_accuracy/inductor_torchbench_training.csv
index dc766eac0..36a646a14 100644
--- a/.github/ci_expected_accuracy/inductor_torchbench_training.csv
+++ b/.github/ci_expected_accuracy/inductor_torchbench_training.csv
@@ -102,5 +102,6 @@ torch_multimodal_clip,pass,pass,pass,eager_fail_to_run,eager_fail_to_run
 tts_angular,pass,eager_fail_to_run,eager_fail_to_run,eager_fail_to_run,eager_fail_to_run
 vgg16,pass,pass,pass,pass,pass
 vision_maskrcnn,pass,pass,pass,eager_fail_to_run,eager_fail_to_run
-yolov3,pass,pass,pass,pass,pass
+# Skip yolov3 for known torchbench issue: https://github.com/intel/torch-xpu-ops/issues/1229
+yolov3,eager_fail_to_run,eager_fail_to_run,eager_fail_to_run,eager_fail_to_run,eager_fail_to_run
 hf_Roberta_base,pass,pass,pass,pass,pass
diff --git a/.github/workflows/pull.yml b/.github/workflows/pull.yml
index 8c0a39c74..fe6e428f5 100644
--- a/.github/workflows/pull.yml
+++ b/.github/workflows/pull.yml
@@ -146,7 +146,7 @@ jobs:
           cp -r ${{ github.workspace }}/../pytorch/inductor_log ${{ github.workspace }}/upload_files
           failed_case=$(grep "Real failed models: *[1-9]" ${{ github.workspace }}/upload_files/summary_accuracy.log |wc -l || true)
           if [ ${failed_case} -ne 0 ];then
-            grep -E "Real failed models: [1-9]|Summary for" ${{ github.workspace }}/summary_accuracy.log
+            grep -E "Real failed models: [1-9]|Summary for" ${{ github.workspace }}/upload_files/summary_accuracy.log
             exit 1
           fi
       - name: Upload Inductor XPU E2E Data