Alan
Results
1
issues of
Alan
Hi, team. I fine tuned with 15 pairs of one document type. (12 pairs as train, 3 as test) After 500 epoches, both of training and eval_loss are nearly zero...