Fix errors after latest PR

Author: volcacius

src/brevitas_examples/llm/llm_quant/mha_layers.py

@@ -3,6 +3,9 @@
 import torch
 from torch import nn
 
+from brevitas.nn.equalized_layer import EqualizedModule
+from brevitas.utils.torch_utils import KwargsForwardHook
+
 
 def attention_mask_handler(
         attention_mask, batch_size, num_heads, query_seq_length, key_value_seq_length):
@@ -53,6 +56,26 @@ def __init__(
             device,
             dtype)
 
+    @property
+    def wrapped_mha(self):
+        mha = self.mha
+        # Workaround for activation equalization for when mha is wrapped
+        # KwargsForwardHook is inserted during act equalization
+        # EqualizedModule is inserted after act equalization
+        if isinstance(mha, KwargsForwardHook):
+            mha = mha.module
+        if isinstance(mha, EqualizedModule):
+            mha = mha.layer
+        return mha
+
+    @property
+    def num_heads(self):
+        return self.wrapped_mha.num_heads
+
+    @property
+    def batch_first(self):
+        return self.wrapped_mha.batch_first
+
     def _load_from_state_dict(
             self, state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys,
             error_msgs):
@@ -134,13 +157,13 @@ def forward(
             key_value_states = hidden_states
         if layer_head_mask is not None:
             raise RuntimeError("layer_head_mask is not supported.")
-        if self.mha.batch_first:
+        if self.batch_first:
             batch_size, query_seq_length = hidden_states.shape[:2]
             key_value_seq_length = key_value_states.shape[1]
         else:
             query_seq_length, batch_size = hidden_states.shape[:2]
             key_value_seq_length = key_value_states.shape[0]
-        num_heads = self.mha.num_heads
+        num_heads = self.num_heads
         attention_mask = attention_mask_handler(
             attention_mask, batch_size, num_heads, query_seq_length, key_value_seq_length)
         attn_output, attn_output_weights = self.mha(

src/brevitas_examples/llm/llm_quant/quantize.py

@@ -268,7 +268,7 @@ def quantize_model(
         'q_scaled_quant': q_scaled_quant,
         'k_transposed_quant': k_transposed_quant,
         'v_quant': v_quant,
-        'out_proj_input_quant': linear_2d_input_quant,
+        'out_proj_input_quant': input_quant,
         'out_proj_weight_quant': weight_quant,
         'out_proj_bias_quant': None,
         'out_proj_output_quant': None,

src/brevitas_examples/llm/main.py

@@ -78,7 +78,7 @@
     default='stats',
     choices=['stats', 'mse'],
     help=
-    'How scales/zero-point are determined. Default: stats (percentile for static, absmax minmax for dynamic).'
+    'How scales/zero-point are determined. Default: stats (percentile for static, absmax or minmax for dynamic).'
 )
 parser.add_argument(
     '--input-scale-precision',
@@ -89,7 +89,7 @@
 parser.add_argument(
     '--input-scale-type',
     type=str,
-    default='float',
+    default='static',
     choices=['static', 'dynamic'],
     help='Whether input scale is a static value or a dynamic value.')
 parser.add_argument(