Fix PyTorch 1.9 compatibility issues.

bindsnet/conversion/conversion.py

@@ -121,10 +121,10 @@ def set_requires_grad(module, value):
 
             extractor2 = FeatureExtractor(module)
             all_activations2 = extractor2.forward(data)
-            for name2, module2 in module.named_children():
+            for name2, module in module.named_children():
                 activations = all_activations2[name2]
 
-                if isinstance(module2, nn.ReLU):
+                if isinstance(module, nn.ReLU):
                     if prev_module is not None:
                         scale_factor = np.percentile(activations.cpu(), percentile)
 
@@ -136,7 +136,7 @@ def set_requires_grad(module, value):
                 elif isinstance(module2, nn.Linear) or isinstance(module2, nn.Conv2d):
                     prev_module = module2
 
-        if isinstance(module2, nn.Linear):
+        if isinstance(module, nn.Linear):
             if prev_module is not None:
                 scale_factor = np.percentile(activations.cpu(), percentile)
                 prev_module.weight *= prev_factor / scale_factor

bindsnet/datasets/collate.py

@@ -8,7 +8,8 @@
 """
 
 import torch
-from torch._six import container_abcs, string_classes, int_classes
+from torch._six import string_classes
+import collections
 
 from torch.utils.data._utils import collate as pytorch_collate
 

bindsnet/pipeline/base_pipeline.py

@@ -2,7 +2,8 @@
 from typing import Tuple, Dict, Any
 
 import torch
-from torch._six import container_abcs, string_classes
+from torch._six import string_classes
+import collections
 
 from ..network import Network
 from ..network.monitors import Monitor

requirements.txt

@@ -1,11 +1,11 @@
 foolbox
-scipy>=1.1.0
-numpy>=1.14.2
-cython>=0.28.5
-torch==1.8.1
-torchvision==0.9.1
+scipy>=1.5.4
+numpy>=1.19.5
+cython>=0.29.5
+torch==1.9.0
+torchvision==0.10.0
 tensorboardX==2.2
-tqdm>=4.19.9
+tqdm>=4.60.0
 setuptools>=39.0.1
 matplotlib>=2.1.0
 gym>=0.10.4

setup.py

@@ -16,20 +16,20 @@
     packages=find_packages(),
     zip_safe=False,
     install_requires=[
-        "numpy>=1.14.2",
-        "torch==1.8.1",
-        "torchvision==0.9.1",
+        "numpy>=1.19.5",
+        "torch==1.9.0",
+        "torchvision==0.10.0",
         "tensorboardX==2.2",
-        "tqdm>=4.19.9",
+        "tqdm>=4.60.0",
         "matplotlib>=2.1.0",
         "gym>=0.10.4",
         "scikit-build>=0.11.1",
         "scikit_image>=0.13.1",
         "scikit_learn>=0.19.1",
         "opencv-python>=3.4.0.12",
-        "pytest>=3.4.0",
-        "scipy>=1.1.0",
-        "cython>=0.28.5",
+        "pytest>=6.2.0",
+        "scipy>=1.5.4",
+        "cython>=0.29.0",
         "pandas>=0.23.4",
     ],
 )

test/conversion/test_conversion.py

@@ -1,3 +1,4 @@
+import torch
 import torch.nn as nn
 import torch.nn.functional as F
 
@@ -24,14 +25,20 @@ def forward(self, x):
         return x
 
 
-def test_conversion():
+def test_conversion_1():
     ann = FullyConnectedNetwork()
     snn = ann_to_snn(ann, input_shape=(784,))
 
 
-def main():
+def test_conversion_2():
+    data = torch.rand(784, 20)
     ann = FullyConnectedNetwork()
-    return ann_to_snn(ann, input_shape=(28, 28))
+    snn = ann_to_snn(ann, data=data, input_shape=(784,))
+
+
+def main():
+    test_conversion_1()
+    test_conversion_2()
 
 
 if __name__ == "__main__":