Question answering for T5 large model

asahi417 / lm-question-generation

Multilingual/multidomain question generation datasets, models, and python library for question generation.

https://www.autoqg.net

MIT License

324 stars 31 forks source link

Question answering for T5 large model #1

Closed cberrioa closed 2 years ago

cberrioa commented 3 years ago

Hi, I have a question regarding the finetuned model "asahi417/question-generation-squad-t5-large". As far as I know, this model was trained using a multitask loss, and question answering is one of these tasks. My question is, is there any specific format to pass the model in order to answer question given a context? Thank you in advance!

asahi417 commented 2 years ago

Hi, sorry for late! The prefix we use for T5 finetuning can be found here. https://github.com/asahi417/t5-question-generation/blob/master/t5qg/lm_t5.py#L19

Also, just want to note that this project is still in progress, so you'll see those model checkpoints updated often.