allow set workspace path for inc quantizer (#987)

## Describe your changes

## Checklist before requesting a review
- [ ] Add unit tests for this change.
- [ ] Make sure all tests can pass.
- [ ] Update documents if necessary.
- [ ] Lint and apply fixes to your code by running `lintrunner -a`
- [ ] Is this a user-facing change? If yes, give a description of this
change to be included in the release notes.
- [ ] Is this PR including examples changes? If yes, please remember to
update [example
documentation](https://github.com/microsoft/Olive/blob/main/docs/source/examples.md)
in a follow-up PR.

## (Optional) Issue link

olive/passes/onnx/inc_quantization.py

@@ -53,6 +53,13 @@
             If users set domain as auto, automatic detection for domain will be executed.
         """,
     ),
+    "workspace": PassConfigParam(
+        type_=str,
+        default_value=None,
+        description="""Workspace for Intel® Neural Compressor quantization where intermediate files and
+            tuning history file are stored.  Default value is:
+            "./nc_workspace/{}/".format(datetime.datetime.now().strftime("%Y-%m-%d_%H-%M-%S"))""",
+    ),
     "recipes": PassConfigParam(
         type_=dict,
         default_value={},
@@ -505,6 +512,12 @@ def _run_for_config(
         eval_func, accuracy_criterion, tuning_criterion = self._set_tuning_config(run_config, data_root)
         weight_only_config = self._set_woq_config(run_config)
 
+        workspace = run_config.pop("workspace", None)
+        if workspace:
+            from neural_compressor import set_workspace
+
+            set_workspace(workspace)
+
         # keys not needed for quantization
         to_delete = [
             "script_dir",

test/unit_test/workflows/test_whisper_workflow.py

@@ -31,7 +31,7 @@ def download_audio_test_data(tmp_path):
 
 
 @pytest.fixture(name="whisper_config")
-def prepare_whisper_config(audio_data):
+def prepare_whisper_config(audio_data, tmp_path):
     return {
         "input_model": {
             "type": "PyTorchModel",
@@ -79,6 +79,9 @@ def prepare_whisper_config(audio_data):
             "inc_dynamic_quantization": {
                 "type": "IncDynamicQuantization",
                 "disable_search": True,
+                "config": {
+                    "workspace": str(tmp_path / "workspace"),
+                },
             },
             "insert_beam_search": {
                 "type": "InsertBeamSearch",
@@ -102,8 +105,8 @@ def prepare_whisper_config(audio_data):
             "evaluate_input_model": False,
             "execution_providers": ["CPUExecutionProvider"],
             "clean_cache": True,
-            "cache_dir": "cache",
-            "output_dir": "models",
+            "cache_dir": str(tmp_path / "cache"),
+            "output_dir": str(tmp_path / "models"),
             "output_name": "whisper_cpu_fp32",
         },
     }