Deprecate weight_only in OVQuantizer (#475)

* Remove deprecated section from documentation

* deprecate weight_only in quantizer

* fix

optimum/intel/openvino/quantization.py

@@ -164,8 +164,11 @@ def quantize(
         if save_directory is None:
             # TODO : can be set to self.model.config.name_or_path for OVModels when not provided
             raise ValueError("`save_directory` needs to be specified")
-
         if weights_only:
+            logger.warning(
+                "Applying weight only quantization using the `OVQuantizer` will be deprecated in the next release of optimum-intel. "
+                "To apply weight only quantization, please set `load_in_8bit=True` when loading your model with `from_pretrained()` or set `--int8` use when exporting your model with the CLI."
+            )
             if calibration_dataset is not None:
                 logger.warning(
                     "`calibration_dataset` was provided but will not be used as `weights_only` is set to `True`."