From 0451cb4c6a4f0d208752ec7f52788a2f62d6cf0d Mon Sep 17 00:00:00 2001
From: aaa123git <wandz19@mails.tsinghua.edu.cn>
Date: Thu, 29 Apr 2021 19:29:09 +0800
Subject: [PATCH] fix bug. len(word_seq) <= 510, since
self.dataloader.pad_batch add two special tokens (#190)
---
convlab2/nlu/jointBERT/multiwoz/nlu.py | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/convlab2/nlu/jointBERT/multiwoz/nlu.py b/convlab2/nlu/jointBERT/multiwoz/nlu.py
index e900f1e..8dd0014 100755
--- a/convlab2/nlu/jointBERT/multiwoz/nlu.py
+++ b/convlab2/nlu/jointBERT/multiwoz/nlu.py
@@ -81,8 +81,8 @@ class BERTNLU(NLU):
da = {}
word_seq, tag_seq, new2ori = self.dataloader.bert_tokenize(ori_word_seq, ori_tag_seq)
- word_seq = word_seq[:512]
- tag_seq = tag_seq[:512]
+ word_seq = word_seq[:510]
+ tag_seq = tag_seq[:510]
batch_data = [[ori_word_seq, ori_tag_seq, intents, da, context_seq,
new2ori, word_seq, self.dataloader.seq_tag2id(tag_seq), self.dataloader.seq_intent2id(intents)]]
--
GitLab