wangxinyu0922 / second_order_sdp Goto Github PK

Hello,
When I was using this repo, I found that the "is_traing=True" in BertVocab still doesn't make the BERT fine-tune due to only the GraphParserNetwork's params is optimized in training. And for current, the "is_traing=True" performance is lower than the "is_traing=False". Could you take a look at it? Thank you!

Hi, I have read your paper, so excellent!
From the README, I noticed there are two ways to get the model you have trained well already(or should be called pretrained model?), and the Labeled F1 is pretty high. So, I downloaded from Google Drive you provide, and consider if the trained model could accomplish the test phase.(Should note that I want to use on my own sentences, i.e., I want to use the trained model to get the SDP parse results on my own sentences), so I tried to transform my sentences to meet your requirement(that is, CONLLU format), and then run the command python3 main.py --save_dir $SAVEDIR run $DATADIR --output_dir results
but, I get a error message, that is, ValueError: Glob of data/SemEval15/DMNew_modified/train.en.dm_modified.conllu yielded no files
I guess there are some data I haven't added into, but I don't want to train a new model, only want to accomplish the test phase first(test on my sentences). So, Do you have any good advice for me , or could you help me use the model you trained right please?
Thank you very much! Looking forward to your reply!

Can you provide an example for using the end to end model for inference?