Question on training dataset
#3
by
aari1995
- opened
Hi again, I was wondering did you train on the whole dataset or just a subset? How long did it take you? Was some pre training involved? And what is an appropriate learning rate, similar to pre training LR?
Thank you!
I was wondering did you train on the whole dataset or just a subset?
Whole
Was some pre training involved?
I used pretrained deberta v3 available on HF
How long did it take you?
About 4/5 hours in one A100 80G GPU
And what is an appropriate learning rate
hyperparameters: https://github.com/urchade/GLiNER/blob/main/config_large.yaml