make threshold and percentage an argument

examples/anomaly_detection_pytorch.py

@@ -130,9 +130,11 @@ def main(args):
             # calculate the surprisal scores for the context target
             scores = -distr.log_prob(context_target / scale)
 
-            # get the top 10% of the scores for each time series of the batch
+            # get the args.top_score_percentage of the scores for each time series of the batch
             top_scores = torch.topk(
-                scores, k=int(scores.shape[1] * 0.1), dim=1
+                scores,
+                k=int(scores.shape[1] * args.top_score_percentage),
+                dim=1,
             )
             # get top scores [B, 10% of context_length]
             top_scores = top_scores.values
@@ -155,7 +157,7 @@ def main(args):
                 # mask out the score where is_anomaly is True
                 score = torch.where(is_anomaly, gpd.loc + 1, score)
                 is_anomaly = torch.where(
-                    is_anomaly, False, gpd.cdf(score) < 0.05
+                    is_anomaly, False, gpd.cdf(score) < args.anomaly_threshold
                 )
                 batch_anomalies.append(is_anomaly)
 
@@ -217,11 +219,23 @@ def main(args):
         "--context_length", type=int, default=None, help="Context length"
     )
     parser.add_argument(
-        "--max_epochs", type=int, default=10, help="Maximum number of epochs"
+        "--max_epochs", type=int, default=30, help="Maximum number of epochs"
     )
     parser.add_argument(
         "--batch_size", type=int, default=32, help="Batch size"
     )
+    parser.add_argument(
+        "--anomaly_threshold",
+        type=float,
+        default=0.05,
+        help="Threshold for anomaly detection",
+    )
+    parser.add_argument(
+        "--top_score_percentage",
+        type=float,
+        default=0.1,
+        help="Percentage of top scores to consider for GPD fitting",
+    )
 
     args = parser.parse_args()