Adjust LayerNormANE bias to match torch.nn.LayerNorm equation and pin…

… torch version

ane_transformers/_version.py

@@ -1 +1 @@
-__version__ = "0.1.1"
+__version__ = "0.1.2"

ane_transformers/huggingface/distilbert.py

@@ -22,6 +22,26 @@
     "coremltools does not support dict outputs. Please set return_dict=False"
 
 
+# Note: torch.nn.LayerNorm and ane_transformers.reference.layer_norm.LayerNormANE
+# apply scale and bias terms in opposite orders. In order to accurately restore a
+# state_dict trained using the former into the the latter, we adjust the bias term
+def correct_for_bias_scale_order_inversion(state_dict, prefix, local_metadata,
+                                           strict, missing_keys,
+                                           unexpected_keys, error_msgs):
+    state_dict[prefix +
+               'bias'] = state_dict[prefix + 'bias'] / state_dict[prefix +
+                                                                  'weight']
+    return state_dict
+
+
+class LayerNormANE(LayerNormANE):
+
+    def __init__(self, *args, **kwargs):
+        super().__init__(*args, **kwargs)
+        self._register_load_state_dict_pre_hook(
+            correct_for_bias_scale_order_inversion)
+
+
 class Embeddings(modeling_distilbert.Embeddings):
     """ Embeddings module optimized for Apple Neural Engine
     """

requirements.txt

@@ -1,4 +1,4 @@
-torch>=1.10.0
+torch>=1.10.0,<=1.11.0
 transformers>=4.18.0
 coremltools>=5.2.0
 yapf

setup.py

@@ -14,7 +14,7 @@
     long_description_content_type='text/markdown',
     author='Apple Inc.',
     install_requires=[
-        "torch>=1.10.0",
+        "torch>=1.10.0,<=1.11.0",
         "coremltools>=5.2.0",
         "transformers>=4.18.0",
         "protobuf>=3.1.0,<=3.20.1",