adlnlp / doc_gcn Goto Github PK

I really enjoy working with your model. Thanks for sharing all materials. Would you please share the code to visualize the predicted results from DocGCN?

I have tried to do a demo of this code but it is not working due to dependencies or missing files :(
Do we have any update?

Hello, I was trying to reproduce your excellent work. But I found out that in the Edge Feature Generation section, the given link "You can get the pre-trained link-prediction model to get the parent-child relation via this [google drive folder]", google drive folder is missing. Could you plz update it?

@adlnlp @yihaoding @siwen26
Hello,
I was trying to reproduce the experiment on FUNSD dataset by following this repo but found some inconsistencies below :

1. The repo's python scripts seems disconnected with missing elements related to final Doc-GCN all feature compilation(appearance, density, syntactic, semantic) and training.
2. The _constituency_parsing_extractor_ seems to be not implemented or defined or imported anywhere.
3. Some of the linked notebooks have google colab drive reference which are not publicly accessible or does on exist with 404 Page Not Found.

For e.g. "our model" here does not exist https://github.com/adlnlp/doc_gcn/blob/main

Can you please help in addressing these issues to fill the voids .

Regards,
Shashank

hello, I spent lot of time trying reproducing the experiment on FUNSD dataset but there are some misconnections.
First of all, the function constituency_parsing_extractor(parse_string) has not been defined, but I am trying to implement it myself.
The second problem is that I found some errors in the function text_density, text_number, char_density and char_number which I think I solved.
The third problem is that in the 3rd notebook you read this pkl file gcn_visual-gcn_char_density-bert_base_cls_test.pkl which has not been defined in the previous notebooks.

I find Doc-GCN a very powerful model and I am trying to solve these issues so I can use your model in other public datasets.
Is there anyone available to help on these issues?