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

for more information, see https://pre-commit.ci

Author: pre-commit-ci[bot]

neural_compressor/torch/algorithms/layer_wise/utils.py

@@ -221,9 +221,7 @@ def load_value(model, param_name, path, device="cpu"):
             os.path.join(path, "model.safetensors"), param_name, prefix=prefix, device=device
         )
     elif len(safetensors_files) >= 2:
-        value = load_tensor_from_safetensors_shard(
-            path, param_name, prefix=prefix, device=device
-        )
+        value = load_tensor_from_safetensors_shard(path, param_name, prefix=prefix, device=device)
     elif "pytorch_model.bin.index.json" in files:
         value = load_tensor_from_shard(path, param_name, prefix)
     else:

neural_compressor/torch/algorithms/weight_only/save_load.py

@@ -22,7 +22,6 @@
 
 import torch
 import transformers
-
 from packaging.version import parse
 
 from neural_compressor.common.utils import AWQ, TEQ, save_config_mapping
@@ -849,8 +848,11 @@ def _init_hf_model(self, model_class, config):
 
             dtype_orig = model_class._set_default_torch_dtype(torch_dtype)
 
-        init_contexts = [no_init_weights(_enable=_fast_init)]  if parse(transformers.__version__) < parse("4.51") else\
-                        [no_init_weights()]
+        init_contexts = (
+            [no_init_weights(_enable=_fast_init)]
+            if parse(transformers.__version__) < parse("4.51")
+            else [no_init_weights()]
+        )
         init_contexts.append(init_empty_weights())
 
         with ContextManagers(init_contexts):

neural_compressor/transformers/models/modeling_auto.py

@@ -38,11 +38,11 @@
 import transformers
 from accelerate import init_empty_weights
 from accelerate.utils import is_xpu_available
+from packaging.version import parse
 from transformers import AutoConfig
 from transformers.configuration_utils import PretrainedConfig
 from transformers.modeling_utils import load_state_dict
 from transformers.utils import has_file, is_safetensors_available
-from packaging.version import parse
 
 from neural_compressor.common.utils import CpuInfo, logger
 from neural_compressor.torch.algorithms.weight_only.modules import INCWeightOnlyLinear
@@ -679,8 +679,11 @@ def load_low_bit(cls, pretrained_model_name_or_path, *model_args, **kwargs):
             quantization_config.weight_dtype = "int4"
             logger.warning("int4 weight_dtype is used, please change the config.json if you don't want to use it.")
 
-        init_contexts = [no_init_weights(_enable=_fast_init)]  if parse(transformers.__version__) < parse("4.51") else\
-                        [no_init_weights()]
+        init_contexts = (
+            [no_init_weights(_enable=_fast_init)]
+            if parse(transformers.__version__) < parse("4.51")
+            else [no_init_weights()]
+        )
         init_contexts.append(init_empty_weights())
 
         with ContextManagers(init_contexts):