What impact does it have on the overall time with this variable set.

awasthiabhijeet / PIE

Fast + Non-Autoregressive Grammatical Error Correction using BERT. Code and Pre-trained models for paper "Parallel Iterative Edit Models for Local Sequence Transduction": www.aclweb.org/anthology/D19-1435.pdf (EMNLP-IJCNLP 2019)

MIT License

228 stars 40 forks source link

What impact does it have on the overall time with this variable set. #16

Closed pidugusundeep closed 4 years ago

pidugusundeep commented 4 years ago

https://github.com/awasthiabhijeet/PIE/blob/2975a6cb3d3ecccddee21d3a0fc094155550dddb/apply_opcode.py#L30

awasthiabhijeet commented 4 years ago

This was implemented to speed up applying opcodes if the test set is very very large. The results in the paper are without using this feature. Also, time for applying opcodes is negligible in comparison to inference time in the PIE model.
I did not spend enough time in testing this feature, as it was not used much. Hence, it is currently set to False.