Neural Coder mod launcher arg: "strategy" to "approach"

neural_coder/__main__.py

@@ -28,8 +28,8 @@ def parse_args():
     parser.add_argument("--opt", type=str, default="",
                         help="optimization feature to enable")
 
-    parser.add_argument("--strategy", type=str, default="static",
-                        help="quantization strategy")
+    parser.add_argument("--approach", type=str, default="static",
+                        help="quantization approach (strategy)")
 
     parser.add_argument('--config', type=str, default="",
                         help='quantization configuration file path')
@@ -53,11 +53,11 @@ def parse_args():
 # optimize on copied script with Neural Coder
 from neural_coder import enable
 if args.opt == "":
-    if args.strategy == "static":
+    if args.approach == "static":
         features=["pytorch_inc_static_quant_fx"]
-    if args.strategy == "static_ipex":
+    if args.approach == "static_ipex":
         features=["pytorch_inc_static_quant_ipex"]
-    if args.strategy == "dynamic":
+    if args.approach == "dynamic":
         features=["pytorch_inc_dynamic_quant"]
 else:
     features=[args.opt]

neural_coder/docs/PythonLauncher.md

@@ -27,7 +27,7 @@ Note: Any modification on the optimized code ```run_glue_optimized.py``` will be
 
 Users can specify which Deep Learning optimization they want to conduct using ```--opt``` argument. The list of supported Deep Learning optimization features can be found [here](SupportMatrix.md).
 
-Note that if specifically optimizing with INT8 quantization by Intel® Neural Compressor, ```--strategy``` argument can be specified with either ```static```, ```static_ipex``` or ```dynamic```. For example, to run INT8 dynamic quantization by Intel® Neural Compressor instead of the default static quantization:
+Note that if specifically optimizing with INT8 quantization by Intel® Neural Compressor, to choose a quantization approach (strategy), ```--approach``` argument can be specified with either ```static```, ```static_ipex``` or ```dynamic```. For example, to run INT8 dynamic quantization by Intel® Neural Compressor instead of the default static quantization:
 ```bash
-python -m neural_coder --strategy dynamic run_glue.py --model_name_or_path bert-base-cased --task_name mrpc --do_eval --output_dir result
+python -m neural_coder --approach dynamic run_glue.py --model_name_or_path bert-base-cased --task_name mrpc --do_eval --output_dir result
 ```