Run black and isort

sinabs/backend/dynapcnn/chips/speck2cmini.py

@@ -2,6 +2,7 @@
 
 import samna
 from samna.speck2cMini.configuration import SpeckConfiguration
+
 from sinabs.backend.dynapcnn.dynapcnn_layer import DynapcnnLayer
 from sinabs.backend.dynapcnn.mapping import LayerConstraints
 

sinabs/backend/dynapcnn/chips/speck2e.py

@@ -2,6 +2,7 @@
 
 import samna
 from samna.speck2e.configuration import SpeckConfiguration
+
 from sinabs.backend.dynapcnn.dynapcnn_layer import DynapcnnLayer
 
 from .dynapcnn import DynapcnnConfigBuilder

sinabs/backend/dynapcnn/chips/speck2f.py

@@ -2,6 +2,7 @@
 
 import samna
 from samna.speck2f.configuration import SpeckConfiguration
+
 from sinabs.backend.dynapcnn.dynapcnn_layer import DynapcnnLayer
 
 from .dynapcnn import DynapcnnConfigBuilder

sinabs/backend/dynapcnn/config_builder.py

@@ -2,10 +2,11 @@
 from typing import Dict, List
 
 import samna
+from samna.dynapcnn.configuration import DynapcnnConfiguration
+
 import sinabs
 import sinabs.backend
 import sinabs.backend.dynapcnn
-from samna.dynapcnn.configuration import DynapcnnConfiguration
 
 from .dvs_layer import DVSLayer
 from .dynapcnn_layer import DynapcnnLayer

sinabs/backend/dynapcnn/discretize.py

@@ -2,10 +2,11 @@
 from typing import Optional, Tuple
 from warnings import warn
 
-import sinabs.layers as sl
 import torch
 import torch.nn as nn
 
+import sinabs.layers as sl
+
 DYNAPCNN_WEIGHT_PRECISION_BITS = 8
 DYNAPCNN_STATE_PRECISION_BITS = 16
 

sinabs/backend/dynapcnn/dvs_layer.py

@@ -1,6 +1,7 @@
 from typing import Optional, Tuple
 
 import torch.nn as nn
+
 from sinabs.layers import SumPool2d
 from sinabs.utils import expand_to_pair
 

sinabs/backend/dynapcnn/dynapcnn_network.py

@@ -61,7 +61,7 @@ def __init__(
                 ... and `dvs_input` is `None`, the `disable_pixel_array` attribute
                     of the layer will not be changed.
                 ... and `dvs_input` is `True`, `disable_pixel_array` will be set
-                    `False`, so that the DVS sensor data is sent to the network. 
+                    `False`, so that the DVS sensor data is sent to the network.
             If no `DVSLayer` is part of `snn`...
                 ... and `dvs_input` is `False` or `None`, no `DVSLayer` will be added
                     and the DVS sensor will not be configured upon deployment.
@@ -138,9 +138,7 @@ def dynapcnn_module(self):
 
     @property
     def exit_layers(self):
-        return [
-            self.all_layers[i] for i in self._dynapcnn_module.get_exit_layers()
-        ]
+        return [self.all_layers[i] for i in self._dynapcnn_module.get_exit_layers()]
 
     @property
     def exit_layer_ids(self):
@@ -525,8 +523,8 @@ def has_dvs_layer(self) -> bool:
         return self.dvs_layer is not None
 
     def zero_grad(self, set_to_none: bool = False) -> None:
-        """ Call `zero_grad` method of each DynapCNN layer
-        
+        """Call `zero_grad` method of each DynapCNN layer
+
         Parameters
         ----------
         - set_to_none (bool): This argument is passed directly to the
@@ -765,9 +763,7 @@ def __repr__(self):
 class DynapcnnCompatibleNetwork(DynapcnnNetwork):
     """Deprecated class, use DynapcnnNetwork instead."""
 
-    def __init__(
-        self, *args, **kwargs
-    ):
+    def __init__(self, *args, **kwargs):
         from warnings import warn
 
         warn(

sinabs/backend/dynapcnn/nir_graph_extractor.py

@@ -211,7 +211,6 @@ def get_dynapcnn_network_module(
                 "destinations": None,
             }
 
-
         # build `DynapcnnLayer` instances from mapper.
         dynapcnn_layers, destination_map, entry_points = (
             construct_dynapcnnlayers_from_mapper(

sinabs/backend/dynapcnn/specksim.py

@@ -4,11 +4,12 @@
 
 import numpy as np
 import samna
-import sinabs.layers as sl
 import torch.nn as nn
 from samna.specksim.nodes import SpecksimConvolutionalFilterNode as ConvFilter
 from samna.specksim.nodes import SpecksimIAFFilterNode as IAFFilter
 from samna.specksim.nodes import SpecksimSumPoolingFilterNode as SumPoolFilter
+
+import sinabs.layers as sl
 from sinabs.backend.dynapcnn import DynapcnnCompatibleNetwork, DynapcnnNetwork
 from sinabs.backend.dynapcnn.dynapcnn_layer import DynapcnnLayer
 

tests/test_dynapcnn/test_discover_device.py

@@ -3,8 +3,9 @@
 
 from sinabs.backend.dynapcnn import io
 
-
 pytest.mark.skip("Not suitable for automated testing. Depends on available devices")
+
+
 def test_is_device_type():
     devices = samna.device.get_all_devices()
     print([io.is_device_type(d, "dynapcnndevkit") for d in devices])

tests/test_dynapcnn/test_dvs_input.py

@@ -130,6 +130,8 @@ def test_dvs_no_pooling(dvs_input):
 
 
 args = product((True, False, None), (True, False))
+
+
 @pytest.mark.parametrize("dvs_input,add_input_layer", args)
 def test_dvs_pooling_2d(dvs_input, add_input_layer):
     # - ANN and SNN generation
@@ -155,7 +157,9 @@ def test_dvs_pooling_2d(dvs_input, add_input_layer):
             spn = DynapcnnNetwork(snn, dvs_input=dvs_input)
 
     # - Compare snn and spn outputs. - Always add DVS so that pooling layer is properly handled
-    spn_float = DynapcnnNetwork(snn, dvs_input=True, discretize=False, input_shape=INPUT_SHAPE)
+    spn_float = DynapcnnNetwork(
+        snn, dvs_input=True, discretize=False, input_shape=INPUT_SHAPE
+    )
     snn_out = snn(input_data).squeeze()
     spn_out = spn_float(input_data).squeeze()
     assert torch.equal(snn_out.detach(), spn_out)

tests/test_dynapcnn/test_speckmini_config_making.py

@@ -70,15 +70,11 @@ def test_auto_mapping():
 
     for test_device in devices:
         # test weights/kernel memory mapping
-        _ = SNN_KERNEL_MEM_TEST.make_config(
-            layer2core_map="auto", device=test_device
-        )
+        _ = SNN_KERNEL_MEM_TEST.make_config(layer2core_map="auto", device=test_device)
         assert SNN_KERNEL_MEM_TEST.layer2core_map == {0: 0, 1: 1, 2: 3, 3: 2, 4: 4}
 
         # test neuron memory mapping
-        _ = SNN_NEURON_MEM_TEST.make_config(
-            layer2core_map="auto", device=test_device
-        )
+        _ = SNN_NEURON_MEM_TEST.make_config(layer2core_map="auto", device=test_device)
         assert SNN_NEURON_MEM_TEST.layer2core_map == {0: 2, 1: 0, 2: 1, 3: 4, 4: 3}
 
 

tests/test_dynapcnn/test_visualizer.py

@@ -44,9 +44,7 @@ def get_demo_dynapcnn_network():
     ann = nn.Sequential(nn.Conv2d(2, 8, (3, 3)), nn.ReLU(), nn.AvgPool2d((2, 2)))
     snn = sinabs.from_model(ann, input_shape=(2, 64, 64), batch_size=1)
 
-    dynapcnn_network = DynapcnnNetwork(
-        snn=snn, input_shape=(2, 64, 64), dvs_input=True
-    )
+    dynapcnn_network = DynapcnnNetwork(snn=snn, input_shape=(2, 64, 64), dvs_input=True)
     return dynapcnn_network
 
 

tests/test_dynapcnnlayer/conftest_dynapcnnlayer.py

@@ -1,10 +1,10 @@
 # author    : Willian Soares Girao
 # contact   : [email protected]
 
-from .model_dummy_1 import expected_output_1, dcnnl_map_1
-from .model_dummy_2 import expected_output_2, dcnnl_map_2
-from .model_dummy_3 import expected_output_3, dcnnl_map_3
-from .model_dummy_4 import expected_output_4, dcnnl_map_4
+from .model_dummy_1 import dcnnl_map_1, expected_output_1
+from .model_dummy_2 import dcnnl_map_2, expected_output_2
+from .model_dummy_3 import dcnnl_map_3, expected_output_3
+from .model_dummy_4 import dcnnl_map_4, expected_output_4
 
 # Args: dcnnl_map, discretize, expected_output
 args_DynapcnnLayer = [

tests/test_dynapcnnlayer/test_dynapcnnlayer.py

@@ -2,12 +2,13 @@
 # contact   : [email protected]
 
 import pytest
-from .conftest_dynapcnnlayer import args_DynapcnnLayer
 
 from sinabs.backend.dynapcnn.dynapcnn_layer_utils import (
     construct_dynapcnnlayers_from_mapper,
 )
 
+from .conftest_dynapcnnlayer import args_DynapcnnLayer
+
 
 @pytest.mark.parametrize(
     "dcnnl_map, discretize, expected_output",

tests/test_dynapcnnnetwork/conftest_dynapcnnnetwork.py

@@ -17,13 +17,12 @@
 from .model_dummy_4 import expected_output as expected_output_4
 from .model_dummy_4 import input_shape as input_shape_4
 from .model_dummy_4 import snn as snn_4
-from .model_dummy_4 import snn as snn_4
 from .model_dummy_seq import (
+    expected_seq_1,
+    expected_seq_2,
     input_shape_seq,
     seq_1,
     seq_2,
-    expected_seq_1,
-    expected_seq_2,
 )
 
 args_DynapcnnNetworkTest = [

tests/test_dynapcnnnetwork/model_dummy_seq.py

@@ -5,7 +5,6 @@
 
 from sinabs.layers import IAFSqueeze, SumPool2d
 
-
 input_shape_seq = (2, 30, 30)
 
 seq_1 = nn.Sequential(

tests/test_dynapcnnnetwork/test_dynapcnnnetwork.py

@@ -3,10 +3,11 @@
 
 import pytest
 import torch
-from .conftest_dynapcnnnetwork import args_DynapcnnNetworkTest
 
 from sinabs.backend.dynapcnn.dynapcnn_network import DynapcnnNetwork
 
+from .conftest_dynapcnnnetwork import args_DynapcnnNetworkTest
+
 
 @pytest.mark.parametrize(
     "snn, input_shape, batch_size, expected_output", args_DynapcnnNetworkTest