Update code

Author: PenghuiCheng

optimum/intel/neural_compressor/__init__.py

@@ -13,7 +13,7 @@
 #  limitations under the License.
 
 from ..utils.import_utils import is_diffusers_available
-from .configuration import INCConfig, WeightOnlyQuantConfig
+from .configuration import INCConfig
 from .modeling_base import (
     INCModel,
     INCModelForMaskedLM,

optimum/intel/neural_compressor/configuration.py

@@ -14,7 +14,6 @@
 
 from typing import Dict, Optional, Union
 
-from intel_extension_for_transformers.transformers.utils import WeightOnlyQuantConfig
 from neural_compressor.config import DistillationConfig, WeightPruningConfig, _BaseQuantizationConfig
 
 from optimum.configuration_utils import BaseConfig
@@ -36,7 +35,7 @@ class INCConfig(BaseConfig):
 
     def __init__(
         self,
-        quantization: Optional[Union[Dict, _BaseQuantizationConfig, WeightOnlyQuantConfig]] = None,
+        quantization=None,
         pruning: Optional[Union[Dict, _BaseQuantizationConfig]] = None,
         distillation: Optional[Union[Dict, _BaseQuantizationConfig]] = None,
         save_onnx_model: bool = False,
@@ -51,7 +50,7 @@ def __init__(
         self.save_onnx_model = save_onnx_model
 
     @staticmethod
-    def _create_quantization_config(config: Union[Dict, _BaseQuantizationConfig, WeightOnlyQuantConfig]):
+    def _create_quantization_config(config):
         # TODO : add activations_dtype and weights_dtype
         if isinstance(config, _BaseQuantizationConfig):
             approach = _quantization_model[config.approach]