Hi, i want to train D3ST with pretrained t5-base, can you give me the clearer way how to train it? (max_length, truncation,...). And if we truncation so do we need to reduce the number of intent asking from the beginning?
For ex : 0=Where to go 1=Go by 1a) on foot 1b) by bike ...... [user] ... [system] ... [user] more than max_length so i think we truncate from the right, but do we need to change the order of 0=Where to go 1=Go by 1a) on foot 1b) by bike ...?
Thanks
Hi, i want to train D3ST with pretrained t5-base, can you give me the clearer way how to train it? (max_length, truncation,...). And if we truncation so do we need to reduce the number of intent asking from the beginning? For ex : 0=Where to go 1=Go by 1a) on foot 1b) by bike ...... [user] ... [system] ... [user] more than max_length so i think we truncate from the right, but do we need to change the order of 0=Where to go 1=Go by 1a) on foot 1b) by bike ...? Thanks