许可证: cc-by-2.0
语言:

• 韩语
• 英语
  任务标签: 文本到文本生成

使用方法

1. 基础模型

from transformers import AutoModelForSeq2SeqLM, AutoTokenizer  

tokenizer = AutoTokenizer.from_pretrained(  
    "wisenut-nlp-team/KoT5",   
    use_auth_token=<个人只读令牌>  
)  
model = AutoModelForSeq2SeqLM.from_pretrained(  
    "wisenut-nlp-team/KoT5",   
    use_auth_token=<个人只读令牌>  
)  

2. 微调模型

摘要生成(summarization)

from transformers import AutoModelForSeq2SeqLM, AutoTokenizer  

tokenizer = AutoTokenizer.from_pretrained(  
    "wisenut-nlp-team/KoT5",   
    revision="summarization",  
    use_auth_token=<个人只读令牌>  
)  
model = AutoModelForSeq2SeqLM.from_pretrained(  
    "wisenut-nlp-team/KoT5",   
    revision="summarization",  
    use_auth_token=<个人只读令牌>  
)  

改写生成(paraphrase generation)

from transformers import AutoModelForSeq2SeqLM, AutoTokenizer  

tokenizer = AutoTokenizer.from_pretrained(  
    "wisenut-nlp-team/KoT5",   
    revision="paraphrase",  
    use_auth_token=<个人只读令牌>  
)  
model = AutoModelForSeq2SeqLM.from_pretrained(  
    "wisenut-nlp-team/KoT5",   
    revision="paraphrase",  
    use_auth_token=<个人只读令牌>  
)