hejunqing
commited on
Commit
•
ac5016c
1
Parent(s):
01ce712
update README
Browse files
README.md
CHANGED
|
@@ -8,9 +8,6 @@ tags:
|
|
| 8 |
- Text Generation
|
| 9 |
- Pretrained Chinese T5-Large model
|
| 10 |
|
| 11 |
-
datasets:
|
| 12 |
-
- CMRC 2018 dev
|
| 13 |
-
|
| 14 |
metrics:
|
| 15 |
- RougeL
|
| 16 |
- BLEU-4
|
|
@@ -60,7 +57,7 @@ pretrain_path = 'IDEA-CCNL/Randeng-T5-784M-QA-Chinese'
|
|
| 60 |
tokenizer=T5Tokenizer.from_pretrained(pretrain_path)
|
| 61 |
model=MT5ForConditionalGeneration.from_pretrained(pretrain_path)
|
| 62 |
|
| 63 |
-
sample={"context":"","question":"","idx":1}
|
| 64 |
plain_text='question:'+sample['question']+'knowledge:'+sample['context'][:self.max_knowledge_length]
|
| 65 |
|
| 66 |
res_prefix=tokenizer.encode('answer'+'<extra_id_0>',add_special_token=False)
|
|
@@ -88,4 +85,4 @@ You can also cite our [website](https://github.com/IDEA-CCNL/Fengshenbang-LM/):
|
|
| 88 |
year={2021},
|
| 89 |
howpublished={\url{https://github.com/IDEA-CCNL/Fengshenbang-LM}},
|
| 90 |
}
|
| 91 |
-
```
|
|
|
|
| 8 |
- Text Generation
|
| 9 |
- Pretrained Chinese T5-Large model
|
| 10 |
|
|
|
|
|
|
|
|
|
|
| 11 |
metrics:
|
| 12 |
- RougeL
|
| 13 |
- BLEU-4
|
|
|
|
| 57 |
tokenizer=T5Tokenizer.from_pretrained(pretrain_path)
|
| 58 |
model=MT5ForConditionalGeneration.from_pretrained(pretrain_path)
|
| 59 |
|
| 60 |
+
sample={"context":"在柏林,胡格诺派教徒创建了两个新的社区:多罗西恩斯塔特和弗里德里希斯塔特。到1700年,这个城市五分之一的人口讲法语。柏林胡格诺派在他们的教堂服务中保留了将近一个世纪的法语。他们最终决定改用德语,以抗议1806-1807年拿破仑占领普鲁士。他们的许多后代都有显赫的地位。成立了几个教会,如弗雷德里夏(丹麦)、柏林、斯德哥尔摩、汉堡、法兰克福、赫尔辛基和埃姆登的教会。","question":"除了多罗西恩斯塔特,柏林还有哪个新的社区?","idx":1}
|
| 61 |
plain_text='question:'+sample['question']+'knowledge:'+sample['context'][:self.max_knowledge_length]
|
| 62 |
|
| 63 |
res_prefix=tokenizer.encode('answer'+'<extra_id_0>',add_special_token=False)
|
|
|
|
| 85 |
year={2021},
|
| 86 |
howpublished={\url{https://github.com/IDEA-CCNL/Fengshenbang-LM}},
|
| 87 |
}
|
| 88 |
+
```
|