JaniceXiong

How to get the keywords/seed words for specific aspect?

1

Hi, thanks for your outstanding work! I saw there are available seed words in the path `data/seeds/`, but I have no idea how to get these seed words if I...

Don't let grobid occupy 8070 port when using docker image

1

Hi, I saw the document that we can map 8070 to 8080 if we want to use grobid on 8080. But in this way, 8070 is still used for grobid....

where is pretrained_model/MASS/dict.txt

1

Hi, thanks for your outstanding work. Here I want to preprocess my own dataset using data_process.sh, but the **srcdict and tgtdict** files are missing in L24 and L25 below. Could...

How to use unlabeled data

7

The paper says "all the left documents are for testing, which are also used as **unlabeled documents during training**", but I did't find the explanation of how to use unlabeled...

Question about candidate summary

2

I see the implementation in 'get_candidate.py': First, we should select **5** most important sentences from BertExt; Then select any 2 or 3 sentences to form a candidate summary. There are...

Question about topic segment representation

1

Hi wendy! @Wendy-Xiao Thanks for your work firstly! I am confused about the generation of topic segment representation. In detail of C, we can use LSTM-Minus to get segment representation....

ZeroDivisionError: float division by zero

9

I met the **"ZeroDivisionError: float division by zero"** when I want to train the model with multi-gpu. And if only 1 gpu, the problem disappear but the training is too...

Train bart.large on custom dataset from the begining

2

Hi, thank you for releasing [trained model](https://drive.google.com/file/d/1BMKhAh2tG5p8THxugZWMPc7NXqwJDHLw/view?usp=sharing). But if I want to train bart.large on my custom dataset from the beginning, and set model_path to [fairseq bart.large](https://dl.fbaipublicfiles.com/fairseq/models/bart.large.tar.gz), raise the exception...