modified files

.gitignore

@@ -11,3 +11,7 @@ venv/
 ENV/
 env.bak/
 venv.bak/
+
+
+IDS17/archive
+IDS17/data

IDS17/CNN1D.py

@@ -24,13 +24,12 @@
 from avalanche.logging import InteractiveLogger, TensorboardLogger, TextLogger
 from avalanche.models import SimpleMLP
 from avalanche.training.plugins import EvaluationPlugin
-from avalanche.training.strategies import GEM
+from avalanche.training.strategies import GEM,EWC 
 from sklearn.manifold import TSNE
 from sklearn.model_selection import StratifiedShuffleSplit, train_test_split
 from torch.nn import CrossEntropyLoss
 from torch.optim import Adam
 from torch.utils.data import Dataset, Subset
-from torchsummary import summary
 from torchvision import transforms
 import sys
 

IDS17/simpleMLP.py

@@ -16,7 +16,7 @@
 import torch
 import torch.nn as nn
 import torchvision
-from avalanche.benchmarks.generators import tensor_scenario
+from avalanche.benchmarks.generators import tensors_benchmark
 from avalanche.evaluation.metrics import (
     ExperienceForgetting,
     StreamConfusionMatrix,
@@ -129,11 +129,13 @@ def get_args(argv):
             temp.loc[:, "70"] = 1
         test_label_dict["cat" + str(i)] = temp
 
-    train_data_x = list(torch.Tensor(train_dict[key].to_numpy()) for key in train_dict)
+    train_data_x = list(torch.Tensor(
+        train_dict[key].to_numpy()) for key in train_dict)
     train_data_y = list(
         torch.Tensor(train_label_dict[key].to_numpy()) for key in train_label_dict
     )
-    test_data_x = list(torch.Tensor(test_dict[key].to_numpy()) for key in test_dict)
+    test_data_x = list(torch.Tensor(
+        test_dict[key].to_numpy()) for key in test_dict)
     test_data_y = list(
         torch.Tensor(test_label_dict[key].to_numpy()) for key in test_label_dict
     )
@@ -166,8 +168,8 @@ def task_ordering(perm: dict):
     """
     final_train_data_x = []
     final_train_data_y = []
-    final_test_data_x = torch.Tensor([])
-    final_test_data_y = torch.Tensor([])
+    final_test_data_x = []
+    final_test_data_y = []
 
     for key, values in perm.items():
         temp_train_data_x = torch.Tensor([])
@@ -176,15 +178,20 @@ def task_ordering(perm: dict):
         temp_test_data_y = torch.Tensor([])
 
         for value in values:
-            temp_train_data_x = torch.cat([temp_train_data_x, train_data_x[value]])
-            temp_train_data_y = torch.cat([temp_train_data_y, train_data_y[value]])
-            temp_test_data_x = torch.cat([temp_test_data_x, test_data_x[value]])
-            temp_test_data_y = torch.cat([temp_test_data_y, test_data_y[value]])
+            temp_train_data_x = torch.cat(
+                [temp_train_data_x, train_data_x[value]])
+            temp_train_data_y = torch.cat(
+                [temp_train_data_y, train_data_y[value]])
+            temp_test_data_x = torch.cat(
+                [temp_test_data_x, test_data_x[value]])
+            temp_test_data_y = torch.cat(
+                [temp_test_data_y, test_data_y[value]])
 
         final_train_data_x.append(temp_train_data_x)
         final_train_data_y.append(temp_train_data_y)
-        final_test_data_x = torch.cat([final_test_data_x, temp_test_data_x])
-        final_test_data_y = torch.cat([final_test_data_y, temp_test_data_y])
+        final_test_data_x.append(temp_test_data_x)
+        final_test_data_y.append(temp_test_data_y)
+
     return final_train_data_x, final_train_data_y, final_test_data_x, final_test_data_y
 
 
@@ -238,12 +245,10 @@ def task_ordering(perm: dict):
 for task_order in range(len(task_order_list)):
     print("Current task order processing ", task_order + 1)
     dataset = task_ordering(task_order_list[task_order])
-
-    generic_scenario = tensor_scenario(
-        train_data_x=dataset[0],
-        train_data_y=dataset[1],
-        test_data_x=dataset[2],
-        test_data_y=dataset[3],
+    print('Len of datasets: ', dataset[0].shape, dataset[2].shape)
+    generic_scenario = tensors_benchmark(
+        train_tensors=[(x, y) for x, y in zip(dataset[0], dataset[1])],
+        test_tensors=[(x, y) for x, y in zip(dataset[2], dataset[3])],
         task_labels=[
             0 for key in task_order_list[task_order].keys()
         ],  # shouldn't provide task ID for inference
@@ -256,20 +261,23 @@ def task_ordering(perm: dict):
 
     # log to text file
     text_logger = TextLogger(
-        open(f"./logs/{cur_time}_simple_mlp_task_0_in_task{task_order}.txt", "w+")
+        open(
+            f"./logs/{cur_time}_simple_mlp_task_0_in_task{task_order}.txt", "w+")
     )
 
     # print to stdout
     interactive_logger = InteractiveLogger()
 
     eval_plugin = EvaluationPlugin(
-        accuracy_metrics(minibatch=True, epoch=True, experience=True, stream=True),
+        accuracy_metrics(minibatch=True, epoch=True,
+                         experience=True, stream=True),
         loss_metrics(minibatch=True, epoch=True, experience=True, stream=True),
         timing_metrics(epoch=True, epoch_running=True),
         ExperienceForgetting(),
         cpu_usage_metrics(experience=True),
         StreamConfusionMatrix(num_classes=2, save_image=False),
-        disk_usage_metrics(minibatch=True, epoch=True, experience=True, stream=True),
+        disk_usage_metrics(minibatch=True, epoch=True,
+                           experience=True, stream=True),
         loggers=[interactive_logger, text_logger, tb_logger],
     )
 
@@ -301,7 +309,7 @@ def task_ordering(perm: dict):
     # TRAINING LOOP
     print("Starting experiment...")
     os.makedirs(
-        os.path.join("weights", f"SimpleMLP_0_in_task_{task_order+1}", exist_ok=True)
+        os.path.join("weights", f"SimpleMLP_0_in_task_{task_order+1}"), exist_ok=True
     )
 
     results = []

IDS17/tSNE.py

@@ -1,7 +1,7 @@
 import pickle
 import warnings
 from datetime import datetime
-from os import path
+import os 
 from typing import Callable, Dict, Iterable
 
 import numpy as np

IDS18/simpleCNN_ClassInc.py

@@ -331,7 +331,7 @@ def forward(self, x):
     print("Starting experiment...")
 
     os.makedirs(
-        os.path.join("weights", f"CNN2D_ClassInc_0inTask{task_order+1}", exist_ok=True)
+        os.path.join("weights", f"CNN2D_ClassInc_0inTask{task_order+1}"), exist_ok=True
     )
 
     results = []

IDS18/simpleCNN_Domain.py

@@ -256,7 +256,7 @@ def forward(self, x):
 # TRAINING LOOP
 print("Starting experiment...")
 
-os.makedirs(os.path.join("weights", f"simpleCNN_Domain", exist_ok=True))
+os.makedirs(os.path.join("weights", f"simpleCNN_Domain"), exist_ok=True)
 
 results = []
 

IDS18/simpleMLP_ClassInc.py

@@ -158,7 +158,7 @@ def get_args(argv):
         pickle.dump(test_data_y, f)
 
 
-def task_ordering(perm: dict):
+def task_ordering(perm):
     """Divides Data into tasks based on the given permutation order
 
     Parameters
@@ -173,8 +173,8 @@ def task_ordering(perm: dict):
     """
     final_train_data_x = []
     final_train_data_y = []
-    final_test_data_x = torch.Tensor([])
-    final_test_data_y = torch.Tensor([])
+    final_test_data_x = []
+    final_test_data_y = []
 
     for key, values in perm.items():
         temp_train_data_x = torch.Tensor([])
@@ -183,17 +183,17 @@ def task_ordering(perm: dict):
         temp_test_data_y = torch.Tensor([])
 
         for value in values:
-            temp_train_data_x = torch.cat([temp_train_data_x, train_data_x[value]])
-            temp_train_data_y = torch.cat([temp_train_data_y, train_data_y[value]])
-            temp_test_data_x = torch.cat([temp_test_data_x, test_data_x[value]])
-            temp_test_data_y = torch.cat([temp_test_data_y, test_data_y[value]])
+            temp_train_data_x = torch.cat([temp_train_data_x, train_data_x[value - 1]])
+            temp_train_data_y = torch.cat([temp_train_data_y, train_data_y[value - 1]])
+            temp_test_data_x = torch.cat([temp_test_data_x, test_data_x[value - 1]])
+            temp_test_data_y = torch.cat([temp_test_data_y, test_data_y[value - 1]])
 
         final_train_data_x.append(temp_train_data_x)
         final_train_data_y.append(temp_train_data_y)
-        final_test_data_x = torch.cat([final_test_data_x, temp_test_data_x])
-        final_test_data_y = torch.cat([final_test_data_y, temp_test_data_y])
-    return final_train_data_x, final_train_data_y, final_test_data_x, final_test_data_y
+        final_test_data_x.append(temp_test_data_x)
+        final_test_data_y.append(temp_test_data_y)
 
+    return final_train_data_x, final_train_data_y, final_test_data_x, final_test_data_y
 
 # Model Creation
 device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
@@ -308,7 +308,7 @@ def task_ordering(perm: dict):
     # TRAINING LOOP
     print("Starting experiment...")
     os.makedirs(
-        os.path.join("weights", f"SimpleMLP_0_in_task_{task_order+1}", exist_ok=True)
+        os.path.join("weights", f"SimpleMLP_0_in_task_{task_order+1}"), exist_ok=True
     )
 
     results = []

IDS18/simpleMLP_Domain.py

@@ -232,7 +232,7 @@ def task_ordering(perm):
 # TRAINING LOOP
 print("Starting experiment...")
 
-os.makedirs(os.path.join("weights", f"simpleMLP_Domain", exist_ok=True))
+os.makedirs(os.path.join("weights", f"simpleMLP_Domain"), exist_ok=True)
 
 results = []
 

KDDCup99/SimpleMLP.py

@@ -49,7 +49,7 @@
 cache_label = "./data/train_data_x.pth"
 # Normal = 0
 # Attack = 1
-if path.exists(cache_label):
+if os.path.exists(cache_label):
 
     with open("./data/train_data_x.pth", "rb") as f:
         train_data_x = pickle.load(f)
@@ -178,17 +178,17 @@ def task_ordering(perm):
     "5": [3, 4, 5, 16],
 }
 
-task_order_list = [perm1]
+task_order_list = [perm]
 
-dataset = task_ordering(task_order_list[task_order])
+dataset = task_ordering(task_order_list[0])
 
 generic_scenario = tensor_scenario(
     train_data_x=dataset[0],
     train_data_y=dataset[1],
     test_data_x=dataset[2],
     test_data_y=dataset[3],
     task_labels=[
-        0 for key in task_order_list[task_order].keys()
+        0 for key in task_order_list[0].keys()
     ],  # shouldn't provide task ID for inference
 )
 
@@ -227,7 +227,7 @@ def task_ordering(perm):
 # TRAINING LOOP
 print("Starting experiment...")
 
-os.makedirs(os.path.join("weights", f"SimpleMLP", exist_ok=True))
+os.makedirs(os.path.join("weights", f"SimpleMLP"), exist_ok=True)
 
 results = []
 i = 1