Datasets:
Tasks:
Text Generation
Modalities:
Text
Formats:
json
Sub-tasks:
language-modeling
Size:
100K - 1M
ArXiv:
Tags:
multilingual
License:
Fine tuning question
#2
by
robbiemu
- opened
did the team ever produce a fine thing dataset as well, or is only this pretraining dataset available?
We have not yet created a finetuning dataset. If necessary, one can aggregate the training instances of each evaluated task mentioned in our paper for fine-tuning.