Skip to content

Latest commit

 

History

History
10 lines (6 loc) · 649 Bytes

README.md

File metadata and controls

10 lines (6 loc) · 649 Bytes

datasize-probing

Reproduced and expanded on experiments from "On the Importance of Data Size in Probing Fine-tuned Models" by Houman Mehrafarin, Sara Rajaee, Mohammad Taher Pilehva. https://arxiv.org/pdf/2203.09627v1.pdf

Use Finetune_GLUE.ipynb to finetune BERT model on GLUE tasks.

Use Finetune_SuperGLUE.ipynb to finetune BERT model on SuperGLUE tasks.

Use SentEval_Probing_notebook.ipynb to perform probing experiments. This code was provided by the author. https://github.com/hmehrafarin/data-size-analysis

Notes about release: Code was simple and mostly uses other libraries so we did not think it warranted a separate PyPi package.