Are you sure you want to delete this task? Once this task is deleted, it cannot be recovered.
Hongzhi (Steve), Chen 704bcaf6dd | 1 year ago | |
---|---|---|
.. | ||
README.md | 3 years ago | |
graphwriter.py | 1 year ago | |
modules.py | 1 year ago | |
opts.py | 1 year ago | |
prepare_data.sh | 4 years ago | |
run.sh | 4 years ago | |
test.sh | 4 years ago | |
train.py | 1 year ago | |
utlis.py | 1 year ago |
In this example we implement the GraphWriter, Text Generation from Knowledge Graphs with Graph Transformers in DGL. And the author's code.
# download data
sh prepare_data.sh
# training
sh run.sh
# testing
sh test.sh
BLEU | METEOR | training time per epoch | |
---|---|---|---|
Author's implementation | 14.3+-1.01 | 18.8+-0.28 | 1970s |
DGL implementation | 14.31+-0.34 | 19.74+-0.69 | 1080s |
We use the author's code for the speed test, and our testbed is V100 GPU.
BLEU | detok BLEU | METEOR | |
---|---|---|---|
greedy, two layers | 13.97 +- 0.40 | 13.78 +- 0.46 | 18.76 +- 0.36 |
beam 4, length penalty 1.0, two layers | 14.66 +- 0.65 | 14.53 +- 0.52 | 19.50 +- 0.49 |
beam 4, length penalty 0.0, two layers | 14.33 +- 0.39 | 14.09 +- 0.39 | 18.63 +- 0.52 |
greedy, six layers | 14.17 +- 0.46 | 14.01 +- 0.51 | 19.18 +- 0.49 |
beam 4, length penalty 1.0, six layers | 14.31 +- 0.34 | 14.35 +- 0.36 | 19.74 +- 0.69 |
beam 4, length penalty 0.0, six layers | 14.40 +- 0.85 | 14.15 +- 0.84 | 18.86 +- 0.78 |
We repeat the experiment five times.
We also provide the output of our implementation on test set together with the reference text.
No Description
Python C++ Jupyter Notebook Cuda Text other
Dear OpenI User
Thank you for your continuous support to the Openl Qizhi Community AI Collaboration Platform. In order to protect your usage rights and ensure network security, we updated the Openl Qizhi Community AI Collaboration Platform Usage Agreement in January 2024. The updated agreement specifies that users are prohibited from using intranet penetration tools. After you click "Agree and continue", you can continue to use our services. Thank you for your cooperation and understanding.
For more agreement content, please refer to the《Openl Qizhi Community AI Collaboration Platform Usage Agreement》