Closed d5423197 closed 2 months ago
Thank you for your interest in our work. The baseline.json
represents the challenge baseline, so it is not our work.
I suggest working directly on the final weights with small modifications to the finetuning configurations (e.g., step2_finetuning/finetuning1.json
).
You may want to change the learning rate to a smaller value.
Please re-open this issue if you have further issues :)
Hello there,
I want to use this repo to train my own dataset. Which config file do you suggest me to start with? The baseline.json?
Thanks,
ZD