diff --git a/run.py b/run.py
index 4026b2fd0..afba5ef93 100644
--- a/run.py
+++ b/run.py
@@ -82,13 +82,15 @@ def run_one_step_with_cudastreams(func, streamcount):
 def printResultSummaryTime(
     result_summary,
     model,
-    metrics_needed=[],
+    metrics_needed=None,
     flops_model_analyzer=None,
     model_flops=None,
     cpu_peak_mem=None,
     mem_device_id=None,
     gpu_peak_mem=None,
 ):
+    if metrics_needed is None:
+        metrics_needed = []
     assert model is not None, "model can not be None."
     if args.device == "cuda":
         gpu_time = np.median(list(map(lambda x: x[0], result_summary)))
@@ -160,11 +162,13 @@ def run_one_step(
     num_iter=10,
     export_metrics_file=None,
     stress=0,
-    metrics_needed=[],
+    metrics_needed=None,
     metrics_gpu_backend=None,
     model_flops=None,
 ):
     # Warm-up `nwarmup` rounds
+    if metrics_needed is None:
+        metrics_needed = []
     for _i in range(nwarmup):
         func()
 
diff --git a/scripts/upload_scribe.py b/scripts/upload_scribe.py
index 392adca69..08cdab230 100644
--- a/scripts/upload_scribe.py
+++ b/scripts/upload_scribe.py
@@ -65,6 +65,7 @@ def upload(self, messages: list):
                     ]
                 ),
             },
+                            timeout=10.0,
         )
         print(r.text)
         r.raise_for_status()
diff --git a/scripts/upload_scribe_v2.py b/scripts/upload_scribe_v2.py
index 441ebd548..9742cb2bf 100644
--- a/scripts/upload_scribe_v2.py
+++ b/scripts/upload_scribe_v2.py
@@ -98,6 +98,7 @@ def upload(self, messages: list):
                     ]
                 ),
             },
+                            timeout=10.0,
         )
         print(r.text)
         r.raise_for_status()
diff --git a/scripts/userbenchmark/upload_scribe.py b/scripts/userbenchmark/upload_scribe.py
index 509768a2e..211c42365 100644
--- a/scripts/userbenchmark/upload_scribe.py
+++ b/scripts/userbenchmark/upload_scribe.py
@@ -64,6 +64,7 @@ def upload(self, messages: list):
                     ]
                 ),
             },
+                            timeout=10.0,
         )
         print(r.text)
         r.raise_for_status()
diff --git a/torchbenchmark/__init__.py b/torchbenchmark/__init__.py
index e15acd828..f6391e039 100644
--- a/torchbenchmark/__init__.py
+++ b/torchbenchmark/__init__.py
@@ -385,8 +385,10 @@ def make_model_instance(
         test: str,
         device: str,
         batch_size: Optional[int] = None,
-        extra_args: List[str] = [],
+        extra_args: List[str] = None,
     ) -> None:
+        if extra_args is None:
+            extra_args = []
         Model = globals()["Model"]
         model = Model(
             test=test, device=device, batch_size=batch_size, extra_args=extra_args
@@ -733,7 +735,7 @@ def get_metadata_from_yaml(path):
     md = None
     if os.path.exists(metadata_path):
         with open(metadata_path, "r") as f:
-            md = yaml.load(f, Loader=yaml.FullLoader)
+            md = yaml.safe_load(f, Loader=yaml.FullLoader)
     return md
 
 
diff --git a/torchbenchmark/_components/model_analyzer/TorchBenchAnalyzer.py b/torchbenchmark/_components/model_analyzer/TorchBenchAnalyzer.py
index 9818798d0..ce6e7b978 100644
--- a/torchbenchmark/_components/model_analyzer/TorchBenchAnalyzer.py
+++ b/torchbenchmark/_components/model_analyzer/TorchBenchAnalyzer.py
@@ -26,12 +26,14 @@ class ModelAnalyzer:
     def __init__(
         self,
         export_metrics_file=None,
-        metrics_needed=[],
+        metrics_needed=None,
         metrics_gpu_backend="nvml",
         cpu_monitored_pid=None,
     ):
         # For debug
         # set_logger(logging.DEBUG)
+        if metrics_needed is None:
+            metrics_needed = []
         set_logger()
         # delay the initialization to start_monitor
         self.gpu_factory = None
diff --git a/torchbenchmark/_components/model_analyzer/dcgm/cpu_monitor.py b/torchbenchmark/_components/model_analyzer/dcgm/cpu_monitor.py
index 8b4eeb69e..2c92f1b2f 100644
--- a/torchbenchmark/_components/model_analyzer/dcgm/cpu_monitor.py
+++ b/torchbenchmark/_components/model_analyzer/dcgm/cpu_monitor.py
@@ -12,7 +12,9 @@ class CPUMonitor(Monitor):
     A CPU monitor that uses psutil to monitor CPU usage
     """
 
-    def __init__(self, frequency, metrics_needed=[], monitored_pid=None):
+    def __init__(self, frequency, metrics_needed=None, monitored_pid=None):
+        if metrics_needed is None:
+            metrics_needed = []
         super().__init__(frequency, metrics_needed)
         # It is a raw record list. [timestamp, cpu_memory_usage, cpu_available_memory]
         self._cpu_records = []
diff --git a/torchbenchmark/canary_models/DALLE2_pytorch/__init__.py b/torchbenchmark/canary_models/DALLE2_pytorch/__init__.py
index bc46969bd..0ff5cbeb9 100644
--- a/torchbenchmark/canary_models/DALLE2_pytorch/__init__.py
+++ b/torchbenchmark/canary_models/DALLE2_pytorch/__init__.py
@@ -21,7 +21,9 @@ class Model(BenchmarkModel):
     DEFAULT_EVAL_BSIZE = 1
     CANNOT_SET_CUSTOM_OPTIMIZER = True
 
-    def __init__(self, test, device, batch_size=None, extra_args=[]):
+    def __init__(self, test, device, batch_size=None, extra_args=None):
+        if extra_args is None:
+            extra_args = []
         super().__init__(
             test=test, device=device, batch_size=batch_size, extra_args=extra_args
         )