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)
Macaulay2MIT
Issues
- 0
about the Edit-factorized BERT Architecture
#33 opened by wangwang110 - 1
Found Garbled in the Synthetic Dataset
#26 opened by KYRIEZX - 2
Running pretrained PIE model gets stuck at INFO:tensorflow:Done running local_init_op.
#30 opened by eedenong - 1
while runing !./multi_round_infer.sh shows error
#17 opened by surianmau - 1
- 0
Small bug in errorify/error.py
#23 opened by kevin-parnow - 0
- 1
which source of correct sentences did you used to make the errorful sentences?
#21 opened by shikha10799 - 1
what's the diff between PIE and LaserTagger ?
#19 opened by Lingogo - 1
usage for last_dot_first_capital?
#20 opened by alexrus - 1
- 10
Bad results trying to train using end_to_end.sh
#18 opened by bexxxxx - 1
- 3
How to fine-tune PIE on customized dataset?
#10 opened by hqsiswiliam - 3
Time taken for multi_round_infer.sh to run?
#15 opened by pidugusundeep - 2
- 2
Inference using own sentences
#14 opened by tuhinjubcse - 4
About Synthetic training
#7 opened by Serenade-J - 0
how to set the number of common deletes?
#13 opened by wangwang110 - 2
Pretrained model bad correction performance?
#11 opened by raff7 - 3
can it use multi gpus to train ?
#8 opened by wangwang110 - 4
Trying to use the pretrained model
#5 opened by MrRobot2211 - 5
Releasing weights
#2 opened by skurzhanskyi - 3
Hi dear authors, I want the script of producing synthetic data, much grateful if you will release it
#1 opened by Serenade-J