[pre-commit.ci] auto fixes from pre-commit hooks

Author: pre-commit-ci[bot]

hls4ml/converters/pytorch/convolution.py

@@ -17,7 +17,7 @@ def parse_conv1d_layer(operation, layer_name, input_names, input_shapes, node, c
     layer['class_name'] = 'Conv1D'
     layer['data_format'] = 'channels_first'  # Pytorch default (can't change)
 
-    if "Quant" in operation:
+    if 'Quant' in operation:
         if class_object.weight_quant.is_quant_enabled:
             width = int(class_object.quant_weight().bit_width)
             scale = class_object.quant_weight().scale.detach().numpy()
@@ -32,8 +32,8 @@ def parse_conv1d_layer(operation, layer_name, input_names, input_shapes, node, c
                 )
             else:
                 raise Exception(
-                    '''Non-power of 2 quantization of weights not supported when injecting brevitas models.
-                    Please used QONNX instead.'''
+                    """Non-power of 2 quantization of weights not supported when injecting brevitas models.
+                    Please used QONNX instead."""
                 )
         else:
             layer['weight_data'] = class_object.weight.data.numpy()
@@ -100,7 +100,7 @@ def parse_conv2d_layer(operation, layer_name, input_names, input_shapes, node, c
     layer['class_name'] = 'Conv2D'
     layer['data_format'] = 'channels_first'  # Pytorch default (can't change)
 
-    if "Quant" in operation:
+    if 'Quant' in operation:
         if class_object.weight_quant.is_quant_enabled:
             width = int(class_object.quant_weight().bit_width)
             scale = class_object.quant_weight().scale.detach().numpy()
@@ -115,8 +115,8 @@ def parse_conv2d_layer(operation, layer_name, input_names, input_shapes, node, c
                 )
             else:
                 raise Exception(
-                    '''Non-power of 2 quantization of weights not supported when injecting brevitas models.
-                    Please used QONNX instead.'''
+                    """Non-power of 2 quantization of weights not supported when injecting brevitas models.
+                    Please used QONNX instead."""
                 )
                 # layer = addQuantizationParameters(layer, class_object.quant_weight(), 'weight')
                 # layer['weight_data'] = class_object.quant_weight().detach().value.numpy()

hls4ml/model/optimizer/passes/brevitas_optimizer.py

@@ -5,7 +5,7 @@
 
 
 class BrevitasInputOutputOptimizer(OptimizerPass):
-    '''Takes nodes parsed from brevitas and inserts Quant nodes into the model if necessary'''
+    """Takes nodes parsed from brevitas and inserts Quant nodes into the model if necessary"""
 
     def match(self, node):
         if ('output_quantization' in node.attributes.keys() and not len(node.attributes['output_quantization']) == 0) or (
@@ -16,7 +16,6 @@ def match(self, node):
             return False
 
     def transform(self, model, node):
-
         # See if Quant layer needs to be added for the output
         if 'output_quantization' in node.attributes.keys() and not len(node.attributes['output_quantization']) == 0:
             print(node.attributes['output_quantization'])
@@ -39,7 +38,6 @@ def transform(self, model, node):
             node.attributes['output_quantization'] = {}
 
         elif 'input_quantization' in node.attributes.keys() and not len(node.attributes['input_quantization']) == 0:
-
             attributes = {}
 
             # Other attributes

test/pytest/test_brevitas_parsing.py

@@ -95,11 +95,11 @@ def test_quantconv1d(backend, io_type):
     if io_type == 'io_stream':
         x = np.ascontiguousarray(x.permute(0, 2, 1))
         config = config_from_pytorch_model(
-            model, (None, n_in, size_in), channels_last_conversion="internal", transpose_outputs=False
+            model, (None, n_in, size_in), channels_last_conversion='internal', transpose_outputs=False
         )
     else:
         config = config_from_pytorch_model(
-            model, (None, n_in, size_in), channels_last_conversion="full", transpose_outputs=True
+            model, (None, n_in, size_in), channels_last_conversion='full', transpose_outputs=True
         )
 
     output_dir = str(test_root_path / f'hls4mlprj_brevitas_conv1d_{backend}_{io_type}')
@@ -156,11 +156,11 @@ def test_quantconv2d(backend, io_type):
     if io_type == 'io_stream':
         x = np.ascontiguousarray(x.permute(0, 2, 3, 1))
         config = config_from_pytorch_model(
-            model, (None, n_in, size_in_height, size_in_width), channels_last_conversion="internal", transpose_outputs=False
+            model, (None, n_in, size_in_height, size_in_width), channels_last_conversion='internal', transpose_outputs=False
         )
     else:
         config = config_from_pytorch_model(
-            model, (None, n_in, size_in_height, size_in_width), channels_last_conversion="full", transpose_outputs=True
+            model, (None, n_in, size_in_height, size_in_width), channels_last_conversion='full', transpose_outputs=True
         )
 
     output_dir = str(test_root_path / f'hls4mlprj_brevitas_conv2d_{backend}_{io_type}')
@@ -270,7 +270,7 @@ def test_pytorch_upsampling1d(data_1d, io_type, backend):
         model,
         (None, in_feat, in_width),
         default_precision='ap_fixed<16,6>',
-        channels_last_conversion="internal",
+        channels_last_conversion='internal',
         transpose_outputs=False,
     )
     odir = str(test_root_path / f'hls4mlprj_pytorch_upsampling_1d_{backend}_{io_type}')
@@ -300,7 +300,7 @@ def test_pytorch_upsampling2d(data_2d, io_type, backend):
         model,
         (in_feat, in_height, in_width),
         default_precision='ap_fixed<16,6>',
-        channels_last_conversion="full",  # With conversion to channels_last
+        channels_last_conversion='full',  # With conversion to channels_last
         transpose_outputs=True,
     )
     odir = str(test_root_path / f'hls4mlprj_pytorch_upsampling_2d_{backend}_{io_type}')
@@ -338,7 +338,7 @@ def test_brevitas_quanteltwiseadd(io_type, backend):
         model,
         [(None, 4, 4), (None, 4, 4)],
         default_precision='ap_fixed<16,6>',
-        channels_last_conversion="off",
+        channels_last_conversion='off',
         transpose_outputs=False,
     )
     odir = str(test_root_path / f'hls4mlprj_brevitas_quanteltwiseadd_{backend}_{io_type}')

test/pytest/test_recurrent_brevitas.py

@@ -55,7 +55,7 @@ def test_rnn(backend, io_type):
     config = config_from_pytorch_model(
         model,
         [(None, 1, 10), (None, 1, 20)],
-        channels_last_conversion="off",
+        channels_last_conversion='off',
         transpose_outputs=False,
         default_precision='fixed<32,16>',
     )