This is an QABot base on 'financezhidao_filter.csv' from https://github.com/SophonPlus/ChineseNlpCorpus
The model are base on hunggingface's transformers https://github.com/huggingface/transformers
For chinese version introduction https://blog.csdn.net/m0_37531129/article/details/102616868
If you want to repeat the repo, pls following steps:
- Please install transformers first refering https://huggingface.co/transformers/installation.html
- Please download finacezhidao.csv from https://github.com/SophonPlus/ChineseNlpCorpus and loaded to preCleanDatafolder then run data clean for finace_zhidao_filter.ipynb to get cleaned training and test data
- Please run bert_update.py using following cmd to get trained model. You can also skip this step, I provid a trained model for you to use run evaluation and test, please download the model from here https://pan.baidu.com/s/1QE1eNr9kd9hDbQK4KKbzhw password: 3uja
- Run bert_update.py --do_predict, you will get the terminal version interaction (it is clumsiness,but this is basic version for test)
- Run python_flask.py to get html version interaction, this is friendly for user (*￣︶￣ *).
Let me know if you have any question or any suggestions!