kanyun-inc fairseq-gec issues

kanyun-inc / fairseq-gec

Source code for paper: Improving Grammatical Error Correction via Pre-Training a Copy-Augmented Architecture with Unlabeled Data

Other

245 stars 67 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

bash train.sh got error

#45 tpackolus opened 1 year ago
1
Inference results using "generate.sh"

#44 Jeiyoon opened 2 years ago
1
I can't run running bash train.sh

#43 aimanmutasem opened 2 years ago
2
How to visualize attention as mentioned in paper for my task(Hindi GEC)?

#42 chayan-dhaddha opened 3 years ago
0
Got an Error while retraining model using ./train.sh

#41 christoph-meyer-horsch opened 3 years ago
1
when single node with multi gpus , it does not run .

#40 hl312 opened 3 years ago
0
running `bash generate.sh` creates model and results folders but there is nothing in them

#39 arikanev opened 4 years ago
0
Why do you need alignfile for our implementation? Please document better.

#38 sabetAI opened 4 years ago
1
How to run this model

#37 jwc19890114 opened 4 years ago
1
update README.md

#36 smiyawaki0820 opened 4 years ago
0
Get raw train/validation text dataset

#35 HelenaHlz closed 4 years ago
0
Could you provide training and validating data?

#34 zixiliuUSC opened 4 years ago
1
The paper mentions that an ensemble model achieved the state of the art results however there is no mention of how the seperate models were trained

#33 Chhokra opened 4 years ago
2
How to train this on a new language?

#32 cppntn opened 4 years ago
0
How do I test the trained model with custom sentences?

#31 Chhokra opened 4 years ago
1
Where is "g.sh"?

#30 jimbozhang closed 4 years ago
2
How should I use pretrained model to process new data

#29 13145314593 opened 4 years ago
4
Hi could you link the final model for inference

#28 MrRobot2211 opened 4 years ago
0
How to make data/train_merge and data/valid?

#27 soyoung97 closed 4 years ago
5
Cross entropy loss giving RuntimeError: CUDA error: device-side assert triggered

#26 Deepayan137 closed 4 years ago
1
Gpu predicts switching to Cpu

#25 1228589545 closed 4 years ago
1
Multi-Task

#24 GrowingZBF closed 4 years ago
4
questions about multi-tasks

#23 linhaow closed 4 years ago
4
About the validation set

#22 xdg988 closed 4 years ago
1
The different number of test data.

#21 EdwardZH closed 4 years ago
0
About the model performance

#20 EdwardZH closed 5 years ago
7
Question about difference of formula of composition_score and copy_score between paper and code

#19 SannyZhou closed 5 years ago
2
Dropout causing negative loss values.

#18 PhilippeMarcotte closed 5 years ago
5
Some problems about run pretrain.sh

#17 cibiluozu closed 5 years ago
2
Question about Single Model Ablation Study on CoNLL 2014 Test Data Set in the paper

#16 ghost closed 5 years ago
5
Questions about training data

#15 luopeixiang closed 4 years ago
3
How to make training data.

#14 ghost closed 5 years ago
5
Got an error while run `sh generate.sh`

#13 tcxdgit opened 5 years ago
7
About pre-train the DAE

#12 wangwang110 closed 5 years ago
1
Details about training the best single model described in the paper

#11 gurunathparasaram closed 5 years ago
1
Fix CUDA_VISIBLE_DEVICES in pretrain.sh

#10 wkwkgg closed 5 years ago
0
Questions regarding checkpoints and interactive.sh

#9 liusiyi641 closed 5 years ago
3
Question about evaluation?

#8 liusiyi641 closed 5 years ago
4
question about the sentences copy task

#7 wangwang110 closed 5 years ago
4
Could you share the fairseq commits id where you init your code?

#6 YuxianMeng closed 5 years ago
2
Question about model parameters, UNK words, loss function and spell error correction system

#5 weiqi94 closed 5 years ago
3
Is negative loss normal?

#4 YuxianMeng closed 5 years ago
2
Update `sh` to `bash` for pre-processing scripts

#3 gurunathparasaram opened 5 years ago
0
Details about the token-level labeling Task

#2 HappyLynn closed 5 years ago
3
Details about training corpus, availability of pre-trained model and data

#1 gurunathparasaram closed 5 years ago
10