The code requires python3. Install dependencies by
pip install -r requirements.txt
If you want to run it on the school server, run the following command to load the some modules first
source scripts/server_env
python3 src/chatbot.py --mode train # train on school server
python3 src/chatbot.py --mode chat # chat on school serverStart the command line interaction with the chat bot.
python src/chatbot.py --mode chatThe model will restore the previously trained weights and continue training up on that. To start training from scratch, please delete all the checkpoints in the checkpoints folder.
python src/chatbot.py --mode trainThe following model parameters can be tuned in config.py
| Parameters | Description |
|---|---|
| NUM_LAYERS | the number of GRU/LSTM layers |
| USE_LSTM | use GRU or LSTM cell |
| USE_DROPOUT | use dropout layer or not |
| DROPOUT_INPUT_KEPP_PROB | [0, 1], input keep probability |
| DROPOUT_OUTPUT_KEEP_PROB | [0, 1], output keep probability |
| DROPOUT_STATE_KEPP_PROB | [0, 1], state dropout is performed on the outgoing states of the cell |
| EMBEDDING_SIZE | the length of the embedding vector for each symbol |
| ATTENTION_HEADS | the number of hidden states that read from the attention state |
| BATCH_SIZE | the size of training data per iteration |
| LR | learning rate of the model |