Limit ITREX version for WOQ (#729)

* remove latest ITREX release compatibility

* update workflow

Author: echarlaix

.github/workflows/test_inc.yml

@@ -34,7 +34,7 @@ jobs:
         pip install py-cpuinfo
         pip install torch==2.2.2 torchvision==0.17.2 torchaudio==2.2.2 --index-url https://download.pytorch.org/whl/cpu
         pip install .[neural-compressor,diffusers,tests]
-        pip install intel-extension-for-transformers
+        pip install intel-extension-for-transformers==1.4.1
         pip install peft
 
     - name: Test with Pytest

optimum/intel/neural_compressor/quantization.py

@@ -79,12 +79,15 @@
 )
 
 
+_ITREX_EXCLUDED_VERSION = "1.4.2"
+
 if is_itrex_available():
     if is_itrex_version("<", ITREX_MINIMUM_VERSION):
         raise ImportError(
             f"Found an incompatible version of `intel-extension-for-transformers`. Found version {_itrex_version}, "
             f"but only version {ITREX_MINIMUM_VERSION} or higher is supported."
         )
+
     from intel_extension_for_transformers.transformers.llm.quantization.utils import convert_to_quantized_model
     from intel_extension_for_transformers.transformers.modeling.modeling_auto import save_low_bit
     from intel_extension_for_transformers.transformers.utils.config import (
@@ -226,6 +229,12 @@ def quantize(
 
         # ITREX Weight Only Quantization
         if not isinstance(quantization_config, PostTrainingQuantConfig):
+            if is_itrex_version("==", _ITREX_EXCLUDED_VERSION):
+                raise ImportError(
+                    f"Found an incompatible version of `intel-extension-for-transformers`. Found version {_itrex_version}, "
+                    f"but {_ITREX_EXCLUDED_VERSION} is not compatible."
+                )
+
             # check neural-compressor version
             if is_neural_compressor_version("<", NEURAL_COMPRESSOR_WEIGHT_ONLY_MINIMUM_VERSION):
                 raise ImportError(