Revert explicit definition of U8 KV-cache

Author: AlexKoff88

optimum/exporters/openvino/__main__.py

@@ -474,9 +474,6 @@ class StoreAttr(object):
         from optimum.intel.openvino.quantization import _weight_only_quantization
 
         _weight_only_quantization(submodel, quantization_config)
-        if "text-generation" in task:
-            submodel.set_rt_info("u8", ["runtime_options", "KV_CACHE_PRECISION"])
-
         compressed_submodel_path = submodel_path.parent / f"{submodel_path.stem}_compressed.xml"
         save_model(submodel, compressed_submodel_path, compress_to_fp16=False)
         del submodel