GitHub - emalgorithm/structured-neural-summarization-replication at 5bbb1850f34029e1ab99b092aac632522bdb9ca3

emalgorithm / structured-neural-summarization-replication Public

Notifications You must be signed in to change notification settings
Fork 3
Star 21

Replication of the paper "Structured Neural Summarization" which uses Graph Neural Networks and Seq2Seq models to summarize natural language and source code.

21 stars 3 forks Branches Tags Activity

Star

Notifications

Branches Tags

Name		Name	Last commit message	Last commit date
Latest commit History 77 Commits
data_processing		data_processing
models		models
training		training
.gitignore		.gitignore
README.md		README.md
graph_pb2.py		graph_pb2.py

Repository files navigation

In order to extract the features from the corpus proto files, run: python data_generation.py

In order to train a model and evaluate a model, run: python train.py --model_name="lstm_gcn_to_lstm_attention" --device=cuda:0 --print_every=10000 --device="cuda" --attention=True --graph=True --iterations=500000 All the possible options when running a model can be seen by running: python train.py --help

A pretrained version of the best performing model (as a state dictionary) can be downloaded at https://drive.google.com/file/d/1fm7hGzr-tziNhUMh8duc8s4j5gWW3uKm/view?usp=sharing

About

Replication of the paper "Structured Neural Summarization" which uses Graph Neural Networks and Seq2Seq models to summarize natural language and source code.

Readme