partial model initialization

layers/tacotron.py

@@ -339,7 +339,6 @@ class Decoder(nn.Module):
 
     def _reshape_memory(self, memory):
         B = memory.shape[0]
-        if memory is not None:
         # Grouping multiple frames if necessary
         if memory.size(-1) == self.memory_dim:
             memory = memory.contiguous()
@@ -370,6 +369,7 @@ class Decoder(nn.Module):
         T = inputs.size(1)
         # Run greedy decoding if memory is None
         greedy = not self.training
+        if memory is not None:
             memory = self._reshape_memory(memory)
         T_decoder = memory.size(0)
         # go frame as zeros matrix

train.py

@@ -401,6 +401,16 @@ def main(args):
     if args.restore_path:
         checkpoint = torch.load(args.restore_path)
         model.load_state_dict(checkpoint['model'])
+        # Partial initialization: if there is a mismatch with new and old layer, it is skipped.
+        # 1. filter out unnecessary keys
+        pretrained_dict = {
+            k: v
+            for k, v in checkpoint['model'].items() if k in model_dict
+        }
+        # 2. overwrite entries in the existing state dict
+        model_dict.update(pretrained_dict)
+        # 3. load the new state dict
+        model.load_state_dict(model_dict)
         if use_cuda:
             model = model.cuda()
             criterion.cuda()