granite speech

Author: Cyrilvallez

src/transformers/models/granite_speech/modeling_granite_speech.py

@@ -330,11 +330,15 @@ def _init_weights(self, module: nn.Module):
             module.weight.data.normal_(mean=0.0, std=std)
             if module.bias is not None:
                 module.bias.data.zero_()
-
         elif isinstance(module, nn.Embedding):
             module.weight.data.normal_(mean=0.0, std=std)
             if module.padding_idx is not None:
                 module.weight.data[module.padding_idx].zero_()
+        elif isinstance(module, (nn.LayerNorm, nn.BatchNorm1d)):
+            module.weight.data.fill_(1.0)
+            module.bias.data.zero_()
+        elif isinstance(module, GraniteSpeechEncoderProjector):
+            module.query.data.normal_()
 
 
 GRANITE_SPEECH_INPUTS_DOCSTRING = r"""