search

marcotcr / checklist

Beyond Accuracy: Behavioral Testing of NLP models with CheckList
MIT License
2.01k stars 204 forks source link

Release of fine-tuned checkpionts #106

Closed yui-ishihara closed 2 years ago

yui-ishihara commented 3 years ago

Hi,

Thank you for this great tool!

Just wondering if you have considered releasing your fine-tuned checkpoints or training scripts (for BERT and RoBERTa) so I could reproduce your experiments (Table 1 through 3 in your original paper).

Thank you. Yui

marcotcr commented 2 years ago

For SQuAD (Table 3), I used this model.

I used the standard huggingface example to train models on GLUE at the time, I'm sure that script is different now. I don't have the checkpoints anymore. I did release the predictions of those models here, so you can run the suites I used for the tables on them and compare to any new models on the same examples