These scripts help you train and evaluate your models on OverLim.
The run_glue.py
script is used to start the training and evaluation, collecting arguments with the help of get_args.py
.
A crude get_results.py
collects the result from the standard output created during training. This can be adjusted to instead look at the evaluation files created during evaluation.
The run_overlim.sh
script shows how to train and evaluate on OverLim with multiple models, collecting the standard output in log-files, that can then be processed by get_results.py
.
You can find some results on our blog: https://kb-labb.github.io/posts/2022-03-16-evaluating-swedish-language-models/