xmudm / basts Goto Github PK

I have difficulty in generating CFG from understand tool. Can you share output of understand tool?? (txt format of any single java program)

hi, i do experiment basts and have some question about valid loss.

in experiment, valid_loss decrease for 100 epochs. and i saw below code.

loss = run_epoch_([val_batches[0]], model,
MultiGPULossCompute(model.generator, criterion, devices=devices, opt=model_opt))

i think opt=model_opt make model train at valid dataset either.

so when i experiments with opt=None , valid loss increased in some epoch ( between 10~20 epoch ).

my question is that..
What method does the performance of paper follow?

感谢您分享您的代码！
我注意到Deepcom中测试集有20000个样例，您预处理后的测试集中有13238个样例，是否一些样例在数据集生成时被筛选掉了？

I'm trying to convert java source codes into CFG-based ast following the readme written on the data preprocessing part.
However, I encountered the problem at executing the "test.pl" file step.
I'm confused about what values the arguments should be set to run "test.pl".
Could you tell me a more specific way to run "test.pl"?

感谢您分享的代码，我从https://drive.google.com/drive/folders/12N-pBzlHhoIgSku7onVPvQrc_JQJw1Q9下载您的数据集后，运行data_preprocess.py文件时，报了如下错误

查看java/train.token.ast数据之后，发现在175406行和175731行存在非float的数据，


且这两行的数据不足512

请问删除这两个数据对最后的结果是否有影响呢？