rasa对话系统实战(1)环境配置+初始
参考文档
(重要)rasa官方文档:
https://rasa.com/docs/rasa/installation
实例中文对话项目参考:
https://www.jianshu.com/p/063d7cd598d3
综合三个对话机器人的融合项目:
https://blog.csdn.net/qq_42189083/article/details/88076128
简单流程参考:
https://paper.yanxishe.com/columnDetail/17542
(重要)rasa简要知识点:
https://blog.csdn.net/ling620/article/details/99845885
MITIE词向量模型详细安装流程:
https://blog.csdn.net/liu765023051/article/details/83107254
Rasa_nlu_chi:入门教程
https://blog.csdn.net/sinat_36226553/article/details/110846622
rasa主要部分为NLU和core:
Rasa NLU 用于聊天机器人中的意图识别和实体提取
“I am looking for a Mexican restaurant in the center of town”
返回结构化数据:
{
"intent": "search_restaurant",
"entities": {
"cuisine" : "Mexican",
"location" : "center"
}
}
环境配置
环境:
python:3.6.4
pycharm: 2017(使用pycharm作为编译器)
VS: 2017
开始配置
在目录E:\1rasa\code下根据rasa官方文档配置环境。
省略文档中过程……
rasa init
或者
rasa init --no-prompt
(如果不用–no-prompt 就会 问很多关于配置的问题)
训练:
rasa train
训练成功的模型存在models文件夹下
only the parts of your model that have changed will be re-trained.
执行:
rasa shell
其他具体命令查看官方文档:https://rasa.com/docs/rasa/command-line-interface
Rasa X 交互界面:
rasa interactive
This will first train a model and then start an interactive shell session. You can then correct your assistants predictions as you talk to it.
http://localhost:5005/visualization.html as soon as the session has started. This diagram can take some time to generate. To skip the visualization, run rasa interactive --skip-visualization.
http://192.168.0.105:5002/login
运行:
rasa run
测试
To evaluate a model on your test data, run:
rasa test
//or
rasa test core
//or
rasa test nlu
数据集分割 rasa data split
To create a train-test split of your NLU training data, run:
rasa data split nlu
前提配置 cmake boost安装
https://blog.csdn.net/caroline2222/article/details/88601373
MITIE安装
中文Rasa_NLU_Chi安装:
训练
1.采用> git clone https://github.com/crownpku/Rasa_NLU_Chi.git 将项目下载到本地;
2. 进入到Rasa_NLU_Chi文件中,打开终端,运行python setup.py install
3.下载训练好的MITIE模型total_word_feature_extractor_chi.dat,分享如下。
链接: https://pan.baidu.com/s/1kZJwLnteouC7e3dKam0PpQ 密码: mdhj
下载后放置位置为Rasa_NLU_Chi/data/total_word_feature_extractor_chi.dat
4.在Rasa_NLU_Chi文件 mkdir mitie, 然后进入到mitie文件下,在终端输入> git clone https://github.com/mit-nlp/MITIE.git 下载MITIE
然后进入到MITIE文件下,在终端输入python setup.py install
5.打开pycharm,打开Rasa_NLU_Chi工程,打开/rasa_nlu/train.py,在Edit Configurations中输入> -c …/sample_configs/config_jieba_mitie_sklearn.yml --data …/data/examples/rasa/demo-rasa_zh.json --path models
然后就可以调试train.py了
————————————————
原文链接:https://blog.csdn.net/Suan2014/article/details/84143330
问题:
1.git安装报错
2.cmake安装后要把路径添加到环境变量中
3.boost安装不成功(找不到bin文件夹。可以编译,无法运行)
4.VS未安装C环境,安装目录在C://(重新安装VS、更换路径)
安装rasa core
https://blog.csdn.net/qq_42189083/article/details/88310407
https://blog.csdn.net/qq_42189083/article/details/88076128