Fix lint error

As the title shows

Co-authored-by: chong-chen <[email protected]>

Author: Jiseong-oh

backends/samsung/_passes/annotate_qparams.py

@@ -9,7 +9,6 @@
 
 import torch
 from executorch.backends.samsung.utils.constants import QuantConstants
-from executorch.backends.samsung.utils.utils import is_graph_input
 from executorch.exir.dialects._ops import ops as exir_ops
 from executorch.exir.pass_base import ExportPass, PassResult
 from torch._export.utils import get_buffer

backends/samsung/builders/__init__.py

@@ -20,9 +20,9 @@
     op_expand_copy,
     op_gelu,
     op_getitem,
+    op_hardsigmoid,
     op_hardswish,
     op_hardtanh,
-    op_hardsigmoid,
     op_layer_norm,
     op_leaky_relu,
     op_linear,

backends/samsung/builders/op_add.py

@@ -12,7 +12,6 @@
     register_node_visitor,
 )
 from executorch.backends.samsung.serialization.enn_graph_schema import EnnGraph
-from executorch.backends.samsung.utils.constants import QuantConstants
 
 
 @register_node_visitor

backends/samsung/builders/op_linear.py

@@ -43,7 +43,7 @@ def define_node(
             all_input_tensors.append(bias_id)
 
         weight_shape = get_shape(weight_node)
-        params = {"in_channels" : weight_shape[1], "out_channels" : weight_shape[0]}
+        params = {"in_channels": weight_shape[1], "out_channels": weight_shape[0]}
         self._update_params_qdtype(node, params)
 
         output_id = self.define_tensor(node, enn_graph, vals_to_ids)

backends/samsung/enn_preprocess.py

@@ -9,14 +9,14 @@
 
 import executorch.backends.samsung.python.PyEnnWrapperAdaptor as PyEnnWrapper
 import torch
+from executorch.backends.samsung._passes.annotate_qparams import AnnotateQparamsPass
 from executorch.backends.samsung._passes.annotate_scalar_parameters import (
     AnnotateScalarParametersPass,
 )
 from executorch.backends.samsung._passes.conv1d_to_conv2d import Conv1dToConv2d
 from executorch.backends.samsung._passes.customized_constant_prop import (
     ConstantPropPass,
 )
-from executorch.backends.samsung._passes.annotate_qparams import AnnotateQparamsPass
 from executorch.backends.samsung._passes.fold_qdq import FoldQDQPass
 from executorch.backends.samsung._passes.insert_qdq import InsertQDQPass
 from executorch.backends.samsung._passes.replace_scalar_ops import ReplaceOpsWithScalar

backends/samsung/serialization/enn_graph_schema.py

@@ -12,7 +12,6 @@
 import numpy as np
 
 import torch
-from executorch.backends.samsung._passes.utils import none_quant_tensor_quant_meta
 from executorch.backends.samsung.builders.utils import DATA_TYPE_STR_MAPPING
 from executorch.backends.samsung.utils.constants import QuantConstants
 from executorch.backends.samsung.utils.utils import quantize_tensor

backends/samsung/utils/export_utils.py

@@ -17,13 +17,11 @@
     DecomposeScaledDotProductAttention,
 )
 from executorch.backends.transforms.remove_clone_ops import RemoveCloneOpsTransform
-from executorch.exir import EdgeCompileConfig, ExportedProgram
+from executorch.exir import EdgeCompileConfig
 from executorch.exir.backend.backend_details import CompileSpec
 from executorch.exir.dialects._ops import ops as exir_ops
 from executorch.exir.pass_manager import PassType
-from executorch.exir.program._program import (
-    to_edge_transform_and_lower,
-)
+from executorch.exir.program._program import to_edge_transform_and_lower
 from torchao.quantization.pt2e.quantize_pt2e import convert_pt2e, prepare_pt2e
 
 

backends/samsung/utils/utils.py

@@ -12,7 +12,6 @@
 from executorch.exir.backend.backend_details import CompileSpec
 from executorch.exir.dialects._ops import ops as exir_ops
 
-from torch._decomp import core_aten_decompositions as torch_core_aten_decompositions
 from torch.export.exported_program import ExportedProgram
 
 

examples/samsung/scripts/wav2letter.py

@@ -168,7 +168,7 @@ def get_dataset(
         help=("Whether to dump all outputs. If not set, we only dump pte."),
         type=bool,
     )
-    
+
     parser.add_argument(
         "-w",
         "--weight",
@@ -196,7 +196,7 @@ def get_dataset(
     instance.vocab_size = 29
     model = instance.get_eager_model().eval()
     if args.weight:
-        weight = torch.load(args.weight)
+        weight = torch.load(args.weight, weights_only=True)
         model.load_state_dict(weight)
     assert args.calibration_number
     if args.dataset: