remove ao folder code (#895)

* remove ao folder code

* update README.md

Author: XiaobingSuper

intel_extension_for_pytorch/__init__.py

@@ -26,7 +26,6 @@
 from . import nn
 from . import jit
 from . import profiler
-from . import ao
 from . import autocast
 
 from .utils.verbose import verbose

intel_extension_for_pytorch/ao/__init__.py

@@ -100,7 +100,7 @@ or define your own qconfig as:
 ```python
 from torch.ao.quantization import MinMaxObserver, PlaceholderObserver, QConfig
 dynamic_qconfig = QConfig(activation = PlaceholderObserver.with_args(dtype=torch.float, compute_dtype=torch.quint8),
-                          weight = MinMaxObserver.with_args(dtype=torch.qint8, qscheme=torch.per_channel_symmetric))
+                          weight = MinMaxObserver.with_args(dtype=torch.qint8, qscheme=torch.per_tensor_symmetric))
 ```
 
 Note: For weight observer, it only supports dtype **torch.qint8**, and the qscheme only can be **torch.per_tensor_symmetric** or **torch.per_channel_symmetric**. For activation observer, it only supports dtype **torch.float**, and the compute_dtype can be **torch.quint8** or **torch.qint8**.

intel_extension_for_pytorch/ao/quantization/__init__.py

@@ -1 +1,2 @@
-from ..ao.quantization import prepare, convert, default_static_qconfig, default_dynamic_qconfig
+from ._quantize import prepare, convert
+from ._qconfig import default_static_qconfig, default_dynamic_qconfig

…on_for_pytorch/ao/quantization/README.md …nsion_for_pytorch/quantization/README.mdintel_extension_for_pytorch/ao/quantization/README.md renamed to intel_extension_for_pytorch/quantization/README.md intel_extension_for_pytorch/ao/quantization/README.md renamed to intel_extension_for_pytorch/quantization/README.md

@@ -8,7 +8,7 @@
 
 import intel_extension_for_pytorch._C as core
 from ._quantize_utils import auto_prepare, auto_convert, copy_prepared_model 
-from ... import nn
+from .. import nn
 
 def prepare(
     model,