Sheng Wang
Forence
·
AI & ML interests
NLP ASR
Organizations
None yet
Forence's activity
lack of digit splitting in slow version of tokenizer
#11 opened 10 months ago
by
Forence

Big difference between the before-cooldown-ckpt and the final checkpoint in the results of downstream tasks?
1
#9 opened about 1 year ago
by
siqi-zz