diegma / graph-2-text Goto Github PK
View Code? Open in Web Editor NEWGraph to sequence implemented in Pytorch combining Graph convolutional networks and opennmt-py
License: MIT License
Graph to sequence implemented in Pytorch combining Graph convolutional networks and opennmt-py
License: MIT License
I'm interest in your paper, thank you for your work.
GCN result in Table 1 of the paper shows 53.5 bleu score. But I can't get the score(highest score is 51.75) after running the project serval times(I have tried to change the batch_size and use gcn gates).
My procedure lists as follow:
(1)Using preprocess.py to preprocess by train-delex files and dev-delex files, e.g train-webnlg-all-delex-src-nodes.txt.
(2)Then I run the train.py. python3 train.py -data data/gcn_exp -save_model data/tmp_ -rnn_size 256 -word_vec_size 256 -layers 1 -epochs 30 -optim adam -learning_rate 0.001 -encoder_type gcn -gcn_num_inputs 256 -gcn_num_units 256 -gcn_in_arcs -gcn_out_arcs -gcn_num_layers 4 -gcn_num_labels 5 -gcn-residual residual -seed 42(43 and 44)
(3)Using translate.py to generate text by test-delex files and relexicalise the text.
(4)Using calculate_bleu_dev.sh to evaluation relexicalised text with only three test-notdelex references(test-all-notdelex-reference0.lex, test-all-notdelex-reference1.lex, test-all-notdelex-reference2.lex).
Is my procedure right? Could you please provide me a exact hyparameters?
Hi, I'm currently working with the WebNLG data and I'd like to know if the texts generated by the models reported in the paper can be made available.
Thanks in advance.
Hi! Thanks a lot for releasing the code.
You said in the ReadMe
There is a preparation step for extracting the node and the edges from the graphs. Instruction for this are in the WebNLG scripts point 1
Could you please point to the link to the script? I couldn't find such script on their website which seems rather complicated.
Thanks a lot in advance!
When I run 'python3 sr11_onmtgcn_input.py -i ../data/srtask11/SR_release1.0/ -t deep -p test', the error occurs:
Traceback (most recent call last):
File "sr11_onmtgcn_input.py", line 19, in
parseService = CoreNLPService()
File "../utils/CoreNLPService.py", line 15, in init
self.server = self.get_server()
File "../utils/CoreNLPService.py", line 18, in get_server
version = jsonrpc.JsonRpc20()
AttributeError: module 'jsonrpc' has no attribute 'JsonRpc20'
I am using python3.6
Hi,
i have problem in processing step. i have downloaded the data on webnlg but i cant found train-webnlg-all-notdelex-src-nodes.txt. where can i get it?
Thanks..
Hi,
Thank you for this great paper and also for releasing the code.
My question is about the preprocessing of other data like dependency parsing, the provided scripts do not handle it I guess. So is it possible to know what is the expected format of the input. Does a Matrix N*N representing the graph structure is enough, or other tricky representations must be added.
Thanks
I am reading your interesting paper and finding this repository link in your paper, but ... why this repository is empty
Hello. I have some trouble with training the model, could you please provide an already trained one?
Thanks.
hi, thanks for releasing the source code. Would you please put the Surface Realization Task (SR11) data here? I searched online and found this website: www.nltg.brighton.ac.uk/research/genchal11/ but I can't find its dataset. And my emails address is [email protected]. Thanks a lot!
Hi, thanks for sharing. May I ask which version of the WebNLP data did you use?
According to the information from https://gitlab.com/shimorina/webnlg-dataset/-/tree/master/, it seems all of the version mismatch the train/dev/test number mentioned in the paper (18102/871/971). Would you mind pointing me out?
File "onmt/io/IO.py", line 244, in get_adj
adj_arc_out[idx_out] = np.array([d, arc_1]) # outgoing arcs
IndexError: index 13440 is out of bounds for axis 0 with size 13440
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.