Update README.md

README.md

@@ -35,7 +35,7 @@ Good at handling NLT tasks, Chinese T5-small， use BertTokenizer and chinese vo
 
 对比T5-small，训练了它的中文版。为了更好适用于中文任务，我们仅使用BertTokenzier，和支持中英文的词表，并且使用了语料库自适应预训练(Corpus-Adaptive Pre-Training, CAPT)技术在悟道语料库(180G版本)继续预训练。预训练目标为破坏span。具体地，我们在预训练阶段中使用了[封神框架](https://github.com/IDEA-CCNL/Fengshenbang-LM/tree/main/fengshen)大概花费了8张A100约24小时。
 
-Compared with T5,-samll we implement its Chinese version. In order to use for chinese tasks, we use BertTokenizer and Chinese vocabulary, and Corpus-Adaptive Pre-Training (CAPT) on the WuDao Corpora (180 GB version). The pretraining objective is span corruption. Specifically, we use the [fengshen framework](https://github.com/IDEA-CCNL/Fengshenbang-LM/tree/main/fengshen) in the pre-training phase which cost about 24 hours with 8 A100 GPUs.
+Compared with T5-samll, we implement its Chinese version. In order to use for chinese tasks, we use BertTokenizer and Chinese vocabulary, and Corpus-Adaptive Pre-Training (CAPT) on the WuDao Corpora (180 GB version). The pretraining objective is span corruption. Specifically, we use the [fengshen framework](https://github.com/IDEA-CCNL/Fengshenbang-LM/tree/main/fengshen) in the pre-training phase which cost about 24 hours with 8 A100 GPUs.
 
 ## 使用 Usage