Merge pull request #687 from BindsNET/hananel

CUDA update and deprecation python 3.8 welcome 3.11

Author: Hananel-Hazan

.github/workflows/pythonpackage.yml

@@ -9,7 +9,7 @@ jobs:
     strategy:
       max-parallel: 4
       matrix:
-        python-version: ["3.8", "3.9", "3.10"]
+        python-version: ["3.9", "3.10", "3.11"]
 
     steps:
     - uses: actions/checkout@v3

bindsnet/datasets/torchvision_wrapper.py

@@ -1,7 +1,7 @@
 from typing import Dict, Optional
 
 import torch
-import torchvision
+from torchvision import datasets as torchDB
 
 from bindsnet.encoding import Encoder, NullEncoder
 
@@ -13,7 +13,7 @@ def create_torchvision_dataset_wrapper(ds_type):
     ``__getitem__``. This applies to all of the datasets inside of ``torchvision``.
     """
     if type(ds_type) == str:
-        ds_type = getattr(torchvision.datasets, ds_type)
+        ds_type = getattr(torchDB, ds_type)
 
     class TorchvisionDatasetWrapper(ds_type):
         __doc__ = (

bindsnet/network/topology.py

@@ -48,6 +48,8 @@ def __init__(
         :param Union[float, torch.Tensor] wmax: Maximum allowed value(s) on the connection weights. Single value, or
             tensor of same size as w
         :param float norm: Total weight per target neuron normalization.
+        :param Union[bool, torch.Tensor] Dales_rule: Whether to enforce Dale's rule. input is boolean tensor in weight shape
+            where True means force zero or positive values and False means force zero or negative values.
         """
         super().__init__()
 
@@ -88,6 +90,12 @@ def __init__(
             **kwargs,
         )
 
+        self.Dales_rule = kwargs.get("Dales_rule", None)
+        if self.Dales_rule is not None:
+            self.Dales_rule = Parameter(
+                torch.as_tensor(self.Dales_rule, dtype=torch.bool), requires_grad=False
+            )
+
     @abstractmethod
     def compute(self, s: torch.Tensor) -> None:
         # language=rst
@@ -117,6 +125,12 @@ def update(self, **kwargs) -> None:
         if mask is not None:
             self.w.masked_fill_(mask, 0)
 
+        if self.Dales_rule is not None:
+            # weight that are negative and should be positive are set to 0
+            self.w[self.w < 0 * self.Dales_rule.to(torch.float)] = 0
+            # weight that are positive and should be negative are set to 0
+            self.w[self.w > 0 * 1 - self.Dales_rule.to(torch.float)] = 0
+
     @abstractmethod
     def reset_state_variables(self) -> None:
         # language=rst