/Chatbot-generate

Automatically train NER&classify models and deploy.

Primary LanguagePythonGNU General Public License v3.0GPL-3.0

chatbot-generate

requirement

python3.6

pytorch

data at dataset

pertrained word vectors at models/glove

configuration file for log database at dataset

Data Format

NER data

<word> <label>

classifier data

label	sentence
<lable> <sentence>

usage:

python main.py or call main(bot_id, ner_file_path, classify_file_path) from main.py

File Orgnization

|- [dir] dataset: data directorys named by bot_id
|- [dir] log
|- [dir] models: pertrained word vectors and well-trained model directorys named by bot_id
|- [dir] ner_evaluation: some result of NER model
|- [dir] service: service interface
|- [dir] service_impl: model loading and forecasting
|- [dir] util: some tools used
|- app.py
|- dataloader.py 
|- loader.py
|- main.py
|- model.py
|- monitor.py
|- train_classify.py
|- train_ner.py
|- utils.py
|- vocab.py

Reference

QAapi

BLSTM-CRF-NER

text-classification