transfer-learning-conv-ai icon indicating copy to clipboard operation
transfer-learning-conv-ai copied to clipboard

Published 20 hours ago • verified publisher icon huggingface

Is it possible to do the auxilary task multiple times for each instance?

Open fabrahman opened this issue 5 years ago • 3 comments

Hi,

Is it possible to define multiple mc_token for each instance and do an auxilary classification for each of them? If yes, can you guide me? Thanks

fabrahman avatar Nov 07 '19 01:11 fabrahman

@thomwolf Will you please guide if possible? Many thanks.

fabrahman avatar Nov 09 '19 22:11 fabrahman

Hi, Hannabrahman. I think you can edit the functions build_input_from_segments, and MODEL_INPUTS in file train.py. And add your classification in the function update

ZHANG45 avatar Jan 06 '20 06:01 ZHANG45

Thanks @ZHANG45. But the GPT2DoubleHeadsModel only accommodates two losses, right? LM and 1 auxiliary multiple choice loss. Should I change the `GPT2DoubleHeadsModel' class as well?

fabrahman avatar Jun 26 '20 17:06 fabrahman