Merge pull request #18 from Vivswan/master

v1.0.0 final

.flake8

@@ -6,6 +6,7 @@ extend-ignore =
     D100, # Missing docstring in public module
     D104, # Missing docstring in public package
     D202, # No blank lines allowed after function docstring
+    D210, # No whitespaces allowed surrounding docstring text
     D401, # First line should be in imperative mood
     R504, # unnecessary variable assignment before return statement
     R505, # unnecessary else after return statement

.gitignore

@@ -3,9 +3,6 @@ __pycache__/
 *.py[cod]
 *$py.class
 
-_run_files/
-_crc_slurm/
-_results/
 _data/
 .idea
 .idea/**

analogvnn/__init__.py

@@ -1,11 +1,12 @@
+"""AnalogVNN: A fully modular framework for modeling and optimizing analog/photonic neural networks."""
+
 import sys
 
 if sys.version_info[:2] >= (3, 8):
     from importlib import metadata
 else:
     import importlib_metadata as metadata  # pragma: no cover
 
-
 __package__ = 'analogvnn'
 __author__ = 'Vivswan Shah (vivswanshah@pitt.edu)'
 

analogvnn/backward/BackwardFunction.py

@@ -1,12 +1,11 @@
 from __future__ import annotations
 
 from abc import ABC
-from typing import Callable
 
 from torch import nn, Tensor
 
 from analogvnn.backward.BackwardModule import BackwardModule
-from analogvnn.utils.common_types import TENSORS
+from analogvnn.utils.common_types import TENSORS, TENSOR_CALLABLE
 
 __all__ = ['BackwardFunction']
 
@@ -15,47 +14,47 @@ class BackwardFunction(BackwardModule, ABC):
     """The backward module that uses a function to compute the backward gradient.
 
     Attributes:
-        _backward_function (Callable): The function used to compute the backward gradient.
+        _backward_function (TENSOR_CALLABLE): The function used to compute the backward gradient.
     """
 
-    _backward_function: Callable
+    _backward_function: TENSOR_CALLABLE
 
-    def __init__(self, backward_function: Callable, layer: nn.Module = None):
+    def __init__(self, backward_function: TENSOR_CALLABLE, layer: nn.Module = None):
         """Initializes the backward module.
 
         Args:
-            backward_function (Callable): The function used to compute the backward gradient.
+            backward_function (TENSOR_CALLABLE): The function used to compute the backward gradient.
             layer (nn.Module): The layer that this backward module is associated with.
         """
 
-        super(BackwardFunction, self).__init__(layer)
+        super().__init__(layer)
         self._backward_function = backward_function
 
     @property
-    def backward_function(self) -> Callable:
+    def backward_function(self) -> TENSOR_CALLABLE:
         """The function used to compute the backward gradient.
 
         Returns:
-            Callable: The function used to compute the backward gradient.
+            TENSOR_CALLABLE: The function used to compute the backward gradient.
         """
 
         return self._backward_function
 
     @backward_function.setter
-    def backward_function(self, backward_function: Callable):
+    def backward_function(self, backward_function: TENSOR_CALLABLE):
         """Sets the function used to compute the backward gradient with.
 
         Args:
-            backward_function (Callable): The function used to compute the backward gradient with.
+            backward_function (TENSOR_CALLABLE): The function used to compute the backward gradient with.
         """
 
         self.set_backward_function(backward_function)
 
-    def set_backward_function(self, backward_function: Callable) -> BackwardFunction:
+    def set_backward_function(self, backward_function: TENSOR_CALLABLE) -> BackwardFunction:
         """Sets the function used to compute the backward gradient with.
 
         Args:
-            backward_function (Callable): The function used to compute the backward gradient with.
+            backward_function (TENSOR_CALLABLE): The function used to compute the backward gradient with.
 
         Returns:
             BackwardFunction: self.

analogvnn/backward/BackwardModule.py

@@ -1,7 +1,7 @@
 from __future__ import annotations
 
 import abc
-from typing import Callable, Any, Optional, Sequence, Tuple, Type
+from typing import Callable, Any, Optional, Tuple, Type
 
 import torch
 from torch import nn, Tensor, autograd
@@ -69,7 +69,7 @@ def backward(ctx: Any, *grad_outputs: Tensor) -> Tuple[None, None, TENSORS]:
             backward_module: BackwardModule = ctx.backward_module
             results = backward_module._call_impl_backward(*grad_outputs)
 
-            if isinstance(results, Sequence):
+            if isinstance(results, (tuple, list)):
                 return (None, None, *results)
 
             return None, None, results
@@ -81,7 +81,7 @@ def __init__(self, layer: nn.Module = None):
             layer (nn.Module): The layer for which the backward gradient is computed.
         """
 
-        super(BackwardModule, self).__init__()
+        super().__init__()
         self._layer = None
         self._set_autograd_backward()
         if not isinstance(self, nn.Module):
@@ -285,4 +285,4 @@ def __getattr__(self, name: str) -> Any:
             return super(BackwardModule, self).__getattr__(name)
         if not str(name).startswith('__') and self._layer is not None and hasattr(self._layer, name):
             return getattr(self._layer, name)
-        raise AttributeError("'{}' object has no attribute '{}'".format(type(self).__name__, name))
+        raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'")

analogvnn/fn/__init__.py

@@ -0,0 +1 @@
+"""Additional functions for analogvnn."""

analogvnn/fn/dirac_delta.py

@@ -2,19 +2,18 @@
 
 from analogvnn.utils.common_types import TENSOR_OPERABLE
 
-__all__ = ['dirac_delta']
+__all__ = ['gaussian_dirac_delta']
 
 
-def dirac_delta(x: TENSOR_OPERABLE, a: TENSOR_OPERABLE = 0.001) -> TENSOR_OPERABLE:
-    """`dirac_delta` takes `x` and returns the Dirac delta function of `x` with standard deviation of `a`.
+def gaussian_dirac_delta(x: TENSOR_OPERABLE, std: TENSOR_OPERABLE = 0.001) -> TENSOR_OPERABLE:
+    """Gaussian Dirac Delta function with standard deviation `std`.
 
     Args:
         x (TENSOR_OPERABLE): Tensor
-        a (TENSOR_OPERABLE): standard deviation.
+        std (TENSOR_OPERABLE): standard deviation.
 
     Returns:
-        TENSOR_OPERABLE: TENSOR_OPERABLE with the same shape as x, but with values equal to the Dirac delta function
-        of x.
+        TENSOR_OPERABLE: TENSOR_OPERABLE with the same shape as x, with values of the Gaussian Dirac Delta function.
     """
 
-    return 1 / (np.abs(a) * np.sqrt(np.pi)) * np.exp(-((x / a) ** 2))
+    return 1 / (np.abs(std) * np.sqrt(np.pi)) * np.exp(-((x / std) ** 2))

analogvnn/graph/AccumulateGrad.py

@@ -30,7 +30,7 @@ def __init__(self, module: Union[nn.Module, Callable]):
             module (Union[nn.Module, Callable]): Module from which to accumulate gradients.
         """
 
-        super(AccumulateGrad, self).__init__()
+        super().__init__()
         self.input_output_connections = {}
         self.module = module
 

analogvnn/graph/AcyclicDirectedGraph.py

@@ -50,7 +50,7 @@ def __init__(self, graph_state: ModelGraphState = None):
             NotImplementedError: If allow_loops is True, since this is not implemented yet.
         """
 
-        super(AcyclicDirectedGraph, self).__init__()
+        super().__init__()
         self.graph = nx.MultiDiGraph()
         self.graph_state = graph_state
         self._is_static = False
@@ -448,3 +448,6 @@ def render(self, *args, real_label: bool = False, **kwargs) -> str:
         """
 
         return to_graphviz_digraph(self.graph, real_label=real_label).render(*args, **kwargs)
+
+    save = render
+    """Alias for render."""

analogvnn/graph/ArgsKwargs.py

@@ -39,7 +39,7 @@ def __init__(self, args=None, kwargs=None):
             kwargs: The keyword arguments.
         """
 
-        super(ArgsKwargs, self).__init__()
+        super().__init__()
         if args is None:
             args = []
         if kwargs is None:
@@ -48,7 +48,7 @@ def __init__(self, args=None, kwargs=None):
         if isinstance(args, tuple):
             args = list(args)
 
-        if not isinstance(args, List):
+        if not isinstance(args, list):
             args = [args]
 
         self.args = args
@@ -79,7 +79,7 @@ def to_args_kwargs_object(cls, outputs: ArgsKwargsInput) -> ArgsKwargs:
             pass
         elif isinstance(outputs, dict):
             outputs = cls(kwargs=outputs)
-        elif isinstance(outputs, tuple) and len(outputs) == 2 and isinstance(outputs[1], dict):
+        elif isinstance(outputs, (tuple, list)) and len(outputs) == 2 and isinstance(outputs[1], dict):
             outputs = cls(args=outputs[0], kwargs=outputs[1])
         else:
             outputs = cls(args=outputs)

analogvnn/graph/ForwardGraph.py

@@ -1,6 +1,6 @@
 from __future__ import annotations
 
-from typing import Dict, Sequence
+from typing import Dict
 
 import torch
 from torch import Tensor
@@ -69,7 +69,7 @@ def calculate(
             ArgsKwargsOutput: Output of the graph
         """
 
-        if not isinstance(inputs, Sequence):
+        if not isinstance(inputs, (tuple, list)):
             inputs = (inputs,)
 
         if not self.graph_state.use_autograd_graph and is_training:

analogvnn/graph/ModelGraph.py

@@ -26,7 +26,7 @@ def __init__(self, use_autograd_graph: bool = False, allow_loops: bool = False):
             allow_loops: If True, the graph is allowed to contain loops.
         """
 
-        super(ModelGraph, self).__init__(use_autograd_graph, allow_loops)
+        super().__init__(use_autograd_graph, allow_loops)
         self.forward_graph = ForwardGraph(self)
         self.backward_graph = BackwardGraph(self)
 

analogvnn/graph/ModelGraphState.py

@@ -46,7 +46,7 @@ def __init__(self, use_autograd_graph: bool = False, allow_loops=False):
             allow_loops: If True, the graph is allowed to contain loops.
         """
 
-        super(ModelGraphState, self).__init__()
+        super().__init__()
         self.allow_loops = allow_loops
         self.use_autograd_graph = use_autograd_graph
 

analogvnn/graph/to_graph_viz_digraph.py

@@ -23,19 +23,14 @@ def to_graphviz_digraph(from_graph: networkx.DiGraph, real_label: bool = False)
         graphviz.Digraph: the converted graph.
 
     Raises:
-        ImportError: if pygraphviz (https://pygraphviz.github.io/) is not available.
         ImportError: if graphviz (https://pygraphviz.github.io/) is not available.
     """
 
-    try:
-        # noinspection PyPackageRequirements
-        import pygraphviz  # noqa: F401
-    except ImportError as e:
-        raise ImportError('requires pygraphviz: https://pygraphviz.github.io/') from e
     try:
         from graphviz import Digraph
     except ImportError as e:
         raise ImportError('requires graphviz: https://pygraphviz.github.io/') from e
+
     strict = networkx.number_of_selfloops(from_graph) == 0 and not from_graph.is_multigraph()
     node_attr = {
         'style': 'filled',
@@ -50,7 +45,6 @@ def to_graphviz_digraph(from_graph: networkx.DiGraph, real_label: bool = False)
         name=from_graph.name,
         strict=strict,
         node_attr=node_attr,
-        graph_attr={'size': '12,12'},
         format='svg'
     )
 

analogvnn/nn/Linear.py

@@ -31,20 +31,19 @@ def forward(self, x: Tensor):
 
         return y
 
-    def backward(self, grad_output: Optional[Tensor], weight: Optional[Tensor] = None) -> Optional[Tensor]:
+    def backward(self, grad_output: Optional[Tensor]) -> Optional[Tensor]:
         """Backward pass of the linear layer.
 
         Args:
             grad_output (Optional[Tensor]): The gradient of the output.
-            weight (Optional[Tensor]): The weight of the layer.
 
         Returns:
             Optional[Tensor]: The gradient of the input.
         """
 
         grad_output = to_matrix(grad_output)
 
-        weight = to_matrix(self.weight if weight is None else weight)
+        weight = to_matrix(self.weight)
         grad_input = grad_output @ weight
 
         self.set_grad_of(self.weight, torch.mm(grad_output.t(), self.inputs))
@@ -77,7 +76,7 @@ def __init__(self, in_features: int, out_features: int, bias: bool = True):
             bias (bool): True if the layer has a bias.
         """
 
-        super(Linear, self).__init__()
+        super().__init__()
         self.in_features = in_features
         self.out_features = out_features
 

analogvnn/nn/activation/ELU.py

@@ -28,7 +28,7 @@ def __init__(self, alpha: float = 1.0507, scale_factor: float = 1.):
             scale_factor (float): the scale factor parameter.
         """
 
-        super(SELU, self).__init__()
+        super().__init__()
         self.alpha = nn.Parameter(torch.tensor(alpha), requires_grad=False)
         self.scale_factor = nn.Parameter(torch.tensor(scale_factor), requires_grad=False)
 
@@ -103,4 +103,4 @@ def __init__(self, alpha: float = 1.0507):
             alpha (float): the alpha parameter.
         """
 
-        super(ELU, self).__init__(alpha=alpha, scale_factor=1.)
+        super().__init__(alpha=alpha, scale_factor=1.)

analogvnn/nn/activation/Identity.py

@@ -23,7 +23,7 @@ def __init__(self, name=None):
             name (str): the name of the activation function.
         """
 
-        super(Identity, self).__init__()
+        super().__init__()
         self.name = name
 
     def extra_repr(self) -> str:

analogvnn/nn/activation/ReLU.py

@@ -28,7 +28,7 @@ def __init__(self, alpha: float):
             alpha (float): the slope of the negative part of the activation function.
         """
 
-        super(PReLU, self).__init__()
+        super().__init__()
         self.alpha = nn.Parameter(torch.tensor(alpha), requires_grad=False)
         self._zero = nn.Parameter(torch.tensor(0), requires_grad=False)
 
@@ -95,7 +95,7 @@ class ReLU(PReLU):
     def __init__(self):
         """Initialize the rectified linear unit (ReLU) activation function."""
 
-        super(ReLU, self).__init__(alpha=0)
+        super().__init__(alpha=0)
 
     @staticmethod
     def initialise(tensor: Tensor) -> Tensor:
@@ -134,4 +134,4 @@ class LeakyReLU(PReLU):
     def __init__(self):
         """Initialize the leaky rectified linear unit (LeakyReLU) activation function."""
 
-        super(LeakyReLU, self).__init__(alpha=0.01)
+        super().__init__(alpha=0.01)

analogvnn/nn/module/Layer.py

@@ -79,7 +79,7 @@ class Layer(nn.Module):
 
     def __init__(self):
         """Initializes the layer."""
-        super(Layer, self).__init__()
+        super().__init__()
         self._inputs = None
         self._outputs = None
         self._backward_module = None
@@ -95,7 +95,7 @@ def __call__(self, *inputs, **kwargs):
         """
 
         self._forward_wrapper(self.forward)
-        outputs = super(Layer, self).__call__(*inputs, **kwargs)
+        outputs = super().__call__(*inputs, **kwargs)
         if self.training:
             self._inputs = ArgsKwargs(args=inputs, kwargs=kwargs)
             self._outputs = outputs