Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/yym6472/ms_pointer_network
Re-implementation of multi-source pointer network.
https://github.com/yym6472/ms_pointer_network
Last synced: about 1 month ago
JSON representation
Re-implementation of multi-source pointer network.
- Host: GitHub
- URL: https://github.com/yym6472/ms_pointer_network
- Owner: yym6472
- Created: 2020-03-13T13:01:18.000Z (almost 5 years ago)
- Default Branch: master
- Last Pushed: 2020-03-25T06:59:51.000Z (over 4 years ago)
- Last Synced: 2024-08-03T09:05:12.891Z (5 months ago)
- Language: Python
- Homepage:
- Size: 12.7 KB
- Stars: 28
- Watchers: 3
- Forks: 22
- Open Issues: 2
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
- StarryDivineSky - yym6472/ms_pointer_network
README
# ms_pointer_network
Re-implementation of multi-source pointer network (from the paper [Multi-Source Pointer Network for Product Title Summarization](https://arxiv.org/pdf/1808.06885.pdf)).
## Requirements
- python - 3.7
- pytorch - 1.4.0
- allennlp - 0.9.0## How to run
### Train model
```
python3 train.py --config_path ./configs/ms_pointer.json --output_dir ./outputs/ms_pointer
```
### Test model
```
python3 test.py --output_dir ./outputs/ms_pointer
```## 数据格式
- tsv文件,第一列为source1的tokens,第二列为source2的tokens,第三列为target的tokens
- 每一列的tokens均使用空格分隔
- target中的词必须在两个source中出现至少一次## TODO
- 目前的实现共用一个词表,是否需要扩展到独立的词表,以解决oov的复制问题?