修复BertEmbedding的bug

fastNLP/embeddings/bert_embedding.py

@@ -294,7 +294,7 @@ class _WordBertModel(nn.Module):
         word_mask = words.ne(self._pad_index)
         seq_len = word_mask.sum(dim=-1)
         batch_word_pieces_length = self.word_pieces_lengths[words]  # batch_size x max_len
-        word_pieces_lengths = batch_word_pieces_length.masked_fill(word_mask, 0).sum(dim=-1)
+        word_pieces_lengths = batch_word_pieces_length.masked_fill(word_mask.eq(0), 0).sum(dim=-1)
         max_word_piece_length = word_pieces_lengths.max().item()
         real_max_word_piece_length = max_word_piece_length  # 表示没有截断的word piece的长度
         if max_word_piece_length+2>self._max_position_embeddings: