Remove HabanaModelRunner D'TOR for mixtral run

vllm/worker/habana_model_runner.py

@@ -16,6 +16,7 @@
 
 import habana_frameworks.torch as htorch
 import torch
+from neural_compressor.torch.quantization import finalize_calibration
 
 from vllm.attention import AttentionMetadata, get_attn_backend
 from vllm.config import (CacheConfig, DeviceConfig, LoadConfig, LoRAConfig,
@@ -1557,7 +1558,6 @@ def prepare_model_input(
                                    virtual_engine=virtual_engine)
 
     def finish_measurements(self):
-        from neural_compressor.torch.quantization import finalize_calibration
         finalize_calibration(self.model.model)
 
     @torch.inference_mode()
@@ -1680,8 +1680,6 @@ def shutdown_inc(self):
         if (model_config := getattr(self, "model_config", None)) and \
                          getattr(model_config, "quantization", None) == 'inc':
             print('inc shutdown start')
-            from neural_compressor.torch.quantization import (
-                finalize_calibration)
             finalize_calibration(self.model.model)
             print('inc shutdown')