5.1 Transformer may have wrong position embed

[JiangHan97]

1. in"class Encoder": enc_outputs = self.src_emb(enc_inputs) + self.pos_emb(torch.LongTensor([[1,2,3,4,0]]))

I think it may be: enc_outputs = self.src_emb(enc_inputs) + self.pos_emb(torch.LongTensor([[0,1,2,3,4]]))

2. in"class Decoder": dec_outputs = self.tgt_emb(dec_inputs) + self.pos_emb(torch.LongTensor([[5,1,2,3,4]]))

I think it may be: dec_outputs = self.tgt_emb(dec_inputs) + self.pos_emb(torch.LongTensor([[0,1,2,3,4]]))