update

Author: Yang Liu

train.py

@@ -33,13 +33,13 @@ def main():
                                        decoder_dim=decoder_dim,
                                        vocab_size=len(word_map),
                                        dropout=dropout)
-        decoder = nn.DataParallel(decoder)
+        # decoder = nn.DataParallel(decoder)
         # decoder = torch.nn.DataParallel(decoder.cuda(), device_ids=[0, 1, 2, 3])
         decoder_optimizer = torch.optim.Adam(params=filter(lambda p: p.requires_grad, decoder.parameters()),
                                              lr=decoder_lr)
         encoder = Encoder()
         encoder.fine_tune(fine_tune_encoder)
-        encoder = nn.DataParallel(encoder)
+        # encoder = nn.DataParallel(encoder)
         # encoder = torch.nn.DataParallel(encoder.cuda(), device_ids=[0, 1, 2, 3])
         encoder_optimizer = torch.optim.Adam(params=filter(lambda p: p.requires_grad, encoder.parameters()),
                                              lr=encoder_lr) if fine_tune_encoder else None