Fix ModuleNotFoundError.

.gitignore

@@ -0,0 +1,8 @@
+__pycache__
+/MA_SNN/DVSGait/data
+/MA_SNN/DVSGestures/data/DvsGesture
+/MA_SNN/DVSGestures/data/train
+/MA_SNN/DVSGestures/data/test
+MA_SNN/DVSGestures/data/DvsGesture.tar.gz
+info.txt
+Attention-SNN.code-workspace

MA_SNN/DVSGestures/Att_SNN_CNN.py

@@ -1,41 +1,42 @@
-import os
-
-import sys
-
-sys.path.append(os.path.dirname("__file__"))
-from DVSGestures.CNN import Att_SNN
-
-rootPath = os.path.abspath(os.path.dirname(__file__))
-rootPath = os.path.split(rootPath)[0]
-sys.path.append(rootPath)
-
-from DVSGestures.CNN import Config
-
-os.environ["CUDA_VISIBLE_DEVICES"] = "4,"
-
-
-class Logger(object):
-    def __init__(self, fileN="Default.log"):
-        self.terminal = sys.stdout
-        self.log = open(fileN, "w")
-
-    def write(self, message):
-        self.terminal.write(message)
-        self.log.write(message)
-
-    def flush(self):
-        pass
-
-
-logPath = Config.configs().recordPath
-if not os.path.exists(logPath):
-    os.makedirs(logPath)
-sys.stdout = Logger(logPath + os.sep + "log_DVS_Gesture_SNN.txt")
-
-
-def main():
-    Att_SNN.main()
-
-
-if __name__ == "__main__":
-    main()
+import os
+
+import sys
+
+# sys.path.append(os.path.dirname("__file__"))
+sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+from DVSGestures.CNN import Att_SNN
+
+rootPath = os.path.abspath(os.path.dirname(__file__))
+rootPath = os.path.split(rootPath)[0]
+sys.path.append(rootPath)
+
+from DVSGestures.CNN import Config
+
+# os.environ["CUDA_VISIBLE_DEVICES"] = "4,"
+
+
+class Logger(object):
+    def __init__(self, fileN="Default.log"):
+        self.terminal = sys.stdout
+        self.log = open(fileN, "w")
+
+    def write(self, message):
+        self.terminal.write(message)
+        self.log.write(message)
+
+    def flush(self):
+        pass
+
+
+logPath = Config.configs().recordPath
+if not os.path.exists(logPath):
+    os.makedirs(logPath)
+sys.stdout = Logger(logPath + os.sep + "log_DVS_Gesture_SNN.txt")
+
+
+def main():
+    Att_SNN.main()
+
+
+if __name__ == "__main__":
+    main()

MA_SNN/DVSGestures/Att_SNN_CNN_SpikingJelly.py

@@ -0,0 +1,42 @@
+import os
+
+import sys
+
+# sys.path.append(os.path.dirname("__file__"))
+sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+from DVSGestures.CNN import Att_SNN_SpikingJelly as Att_SNN
+
+rootPath = os.path.abspath(os.path.dirname(__file__))
+rootPath = os.path.split(rootPath)[0]
+sys.path.append(rootPath)
+
+from DVSGestures.CNN import Config
+
+# os.environ["CUDA_VISIBLE_DEVICES"] = "0"
+
+
+class Logger(object):
+    def __init__(self, fileN="Default.log"):
+        self.terminal = sys.stdout
+        self.log = open(fileN, "w")
+
+    def write(self, message):
+        self.terminal.write(message)
+        self.log.write(message)
+
+    def flush(self):
+        pass
+
+
+logPath = Config.configs().recordPath
+if not os.path.exists(logPath):
+    os.makedirs(logPath)
+sys.stdout = Logger(logPath + os.sep + "log_DVS_Gesture_SNN.txt")
+
+
+def main():
+    Att_SNN.main()
+
+
+if __name__ == "__main__":
+    main()

MA_SNN/DVSGestures/CNN/Att_SNN_SpikingJelly.py

@@ -0,0 +1,44 @@
+import torch
+from utils import util
+from DVSGestures.DVS_Gesture_utils.dataset import create_data
+from DVSGestures.CNN.Networks.Att_SNN_SpikingJelly import create_net
+from DVSGestures.CNN.Config import configs
+from DVSGestures.DVS_Gesture_utils.process import process
+from DVSGestures.DVS_Gesture_utils.save import save_csv
+
+
+def main():
+
+    config = configs()
+    config.device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+    print(config.device)
+
+    config.device_ids = range(torch.cuda.device_count())
+    print(config.device_ids)
+
+    config.name = (
+        config.attention
+        + "_SNN(CNN)-DVS-Gesture_dt="
+        + str(config.dt)
+        + "ms"
+        + "_T="
+        + str(config.T)
+    )
+    config.modelNames = config.name + ".t7"
+    config.recordNames = config.name + ".csv"
+
+    print(config)
+
+    create_data(config=config)
+
+    create_net(config=config)
+
+    print(config.model)
+
+    print(util.get_parameter_number(config.model))
+
+    process(config=config)
+
+    print("best acc:", config.best_acc, "best_epoch:", config.best_epoch)
+
+    save_csv(config=config)

MA_SNN/DVSGestures/CNN/Config.py

@@ -1,13 +1,14 @@
 import os, torch
 import torch.nn as nn
+from spikingjelly.activation_based.neuron import surrogate
 
 
 class configs(object):
     def __init__(self):
-        self.dt = 25
+        self.dt = 15
         self.T = 60
 
-        self.attention = "no"
+        self.attention = "TCSA"
         self.c_ratio=8
         self.t_ratio=5
         self.epoch = 0
@@ -16,7 +17,7 @@ def __init__(self):
         self.pretrained_path = None
 
         self.batch_size = 128
-        self.batch_size_test = 128
+        self.batch_size_test = 28
 
         # None  'kaiming'  'xavier'
         self.init_method = None
@@ -37,18 +38,24 @@ def __init__(self):
         self.interval_scaling = False
 
         # network
-        self.beta = 0
+        self.beta = 0.
         self.alpha = 0.3
-        self.Vreset = 0
+        self.Vreset = 0.
         self.Vthres = 0.3
         self.reduction = 16
         self.T_extend_Conv = False
         self.T_extend_BN = False
         self.h_conv = False
+        self.step_mode = "m"
+        # self.surrogate_function = surrogate.Sigmoid()
+        self.surrogate_function = surrogate.LeakyKReLU()
+        self.backend = "cupy"
+        # Old parameters
         self.mem_act = torch.relu
         self.mode_select = "spike"
         self.TR_model = "NTR"
 
+
         # BatchNorm
         self.track_running_stats = True
 

MA_SNN/DVSGestures/CNN/Networks/Att_SNN.py

@@ -195,18 +195,24 @@ def __init__(
             )
 
         def forward(self, input):
+            # print("input: ",input.shape)
             b, t, _, _, _ = input.size()
             outputs = input
 
             outputs = self.convAttLIF0(outputs)
+            # print("convAttLIF0: ",outputs.shape)
             outputs = self.convAttLIF1(outputs)
+            # print("convAttLIF1: ",outputs.shape)
             outputs = self.convAttLIF2(outputs)
+            # print("convAttLIF1: ",outputs.shape)
 
             outputs = outputs.reshape(b, t, -1)
-
+            # print("fc_input: ",outputs.shape)
             outputs = self.FC0(outputs)
+            # print("FC0: ",outputs.shape)
 
             outputs = self.FC1(outputs)
+            # print("FC1: ",outputs.shape)
             outputs = torch.sum(outputs, dim=1)
             outputs = outputs / t