https://github.com/olenet/nips_convai2_resource
Maintain resources for NIPS 2018 Competition ConvAi2, including dataset url, papers, etc
https://github.com/olenet/nips_convai2_resource
awesome chatbot chinese competition convai2 conversation nips nips-challenge nips2017
Last synced: about 2 hours ago
JSON representation
Maintain resources for NIPS 2018 Competition ConvAi2, including dataset url, papers, etc
- Host: GitHub
- URL: https://github.com/olenet/nips_convai2_resource
- Owner: OleNet
- Created: 2018-05-05T03:06:22.000Z (about 8 years ago)
- Default Branch: master
- Last Pushed: 2018-05-06T08:50:15.000Z (about 8 years ago)
- Last Synced: 2024-05-20T02:52:50.330Z (about 2 years ago)
- Topics: awesome, chatbot, chinese, competition, convai2, conversation, nips, nips-challenge, nips2017
- Homepage:
- Size: 25.4 KB
- Stars: 6
- Watchers: 3
- Forks: 2
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Resources for NIPS 2018 Competition The Conversational Intelligence Challenge 2 (ConvAI2)
## ParlAI
安装FAIR实验室的工具[ParlAI](https://github.com/facebookresearch/ParlAI#installing-parlai)
## Dataset
Official train dataset download [link](https://s3.amazonaws.com/fair-data/parlai/convai2/convai2.tar.gz)
Or
中国用户下载amazonaws的资源比较慢，这里有一个百度云盘[convai2.tar.gz](https://pan.baidu.com/s/1Akz9PpbNvlvVMtp9NxXBDA)的镜像,
下载的convai2.tar.gz 放在如下路径:ParlAI/data/ConvAI2/convai2.tar.gz, 注意不需要解压。
程序自动处理后，ParlAI/data/ConvAI2/目录中会包含如下文件：
|id| `ls ParlAI/data/ConvAI2/` |
|--|-----------------------------------|
|1| train_both_original_no_cands.txt |
|2| train_both_original.txt |
|3| train_both_revised_no_cands.txt |
|| ... ... |
|27| valid_self_revised_no_cands.txt |
|28| valid_self_revised.txt |
文件名也代表了数据的组织方式
按照是否含有self/other 的 persona(画像)组织
| both | none | other | self | 说明 |
|------|------|-------|------|---------------|
| O | X | X | X | 两者的persona都有 |
| X | O | X | X | 两者的persona都没有 |
| X | X | O | X | 只有对方的persona |
| X | X | X | O | 只有自己的persona |
按照修正前/后的persona组织
| original | revise | 说明 |
|----------|--------|-----------------|
| O | X | 原始的persona产出的数据 |
| X | O | 修正的persona产出的数据 |
按照有无候选组织
| no_cands | 说明 |
|----------|-----|
| O | 无候选 |
## Baseline models
因为ParlAI的baseline模块依赖了很多预训练资源，存在aws上不方便中国用户下载，所以为了方便，把这些与训练的资源已经在[百度云](https://pan.baidu.com/s/1PwsbFJo0FIvEhmiu4T4AIw)上。
下载之后, 在ParlAI/data/（没有的话创建一个即可）文件夹下解压model.tar.gz，`tar zxvf model.tar.gz`.
## Run Example
run一个baseline脚本，测试是否上述数据准备成功：
`cd ParlAI; python projects/convai2/baselines/kvmemnn/eval_f1.py`
## Papers
[A Persona-Based Neural Conversation Model](https://arxiv.org/abs/1603.06155)
[A Neural Conversational Model](https://arxiv.org/abs/1506.05869)
[A Diversity-Promoting Objective Function for Neural Conversation Models](https://arxiv.org/abs/1510.03055)
[Personalizing Dialogue Agents: I have a dog, do you have pets too?](https://arxiv.org/abs/1801.07243)