https://github.com/justdvnsh/dbot

My chatbot
https://github.com/justdvnsh/dbot

chatbot dbot decoder deep-learning deep-neural-networks encoder encoder-decoder-model gru lstm machine-learning machine-learning-algorithms rnn rnn-tensorflow seq2seq

Last synced: 11 months ago
JSON representation

My chatbot

• Host: GitHub
• URL: https://github.com/justdvnsh/dbot
• Owner: justdvnsh
• Created: 2018-12-17T10:49:04.000Z (about 7 years ago)
• Default Branch: master
• Last Pushed: 2019-01-21T14:35:39.000Z (about 7 years ago)
• Last Synced: 2025-02-07T08:31:17.359Z (about 1 year ago)
• Topics: chatbot, dbot, decoder, deep-learning, deep-neural-networks, encoder, encoder-decoder-model, gru, lstm, machine-learning, machine-learning-algorithms, rnn, rnn-tensorflow, seq2seq
• Language: Jupyter Notebook
• Size: 9.29 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# DBot

My personal chatbot. The website where you can try the chatbot will be released quite soon, as the model is made, where you can talk to the chatbot . 

Don't expect a chatbot as complicated as Assistant or Siri, yet it is a powerful chatbot, you can surely have a good time chatting with. Also, this 

chatbot is trained on movie dialouges , so expect quite filmy dialogues .

# Current Progress

Since I am in need of a personal GPU, I have been training the models on the free GPU sessions on Kaggle and colab. Now , since the free sessions have limits

I could not train the model on massive datasets, but I managed to train the models on 1000 sentences of the [CORNELL-MOVIE-DATASET](https://www.kaggle.com/) dataset,

since it contains both the text and summary. I also tried the model to run on the BBC news dataset, but , unfortunately the free session could not manage the massive size of the dataset.

Thus I trained the model for 10 epochs , using the GRU cell and seq2seq with attention mechanism . So, being said that , these are the results I have got.

```python

while True:

    question = input("You: ")

    if question.lower() == 'goodbye':

        break

    else:

        result = answer(question, 

                        encoder, 

                        decoder, 

                        inp_lang, 

                        targ_lang, 

                        max_length_inp, 

                        max_length_targ)  

    print('Chatbot: ', result)

```

    You: how are you?

    Chatbot:  it doesn t tell me about back i was it back when it .  

    You: what's your name again ?

    Chatbot:  okay . what just worlds n . . . maybe you know you are eddie .  

    You: okay, so you are eddie

    Chatbot:  oh , eddie .  

    You: hi

    Chatbot:  tamina could know ! i supposed to doin me done , sweetheart ?  

    You: is your name cameron ?

    Chatbot:  i ride for , just for your son , we questioned you wanna have , tabloids chosen for a good . like his get . i used out with me wrong . not s a hell , too . and by your america .  

    You: how do you get your hair to look like that ?

    Chatbot:  again .  

    You: yes

    Chatbot:  does . rights stupid ways run sinned .  

    You: how do you get your hair like that ?

    Chatbot:  william son . i thought  

    You: goodbye

    

	

# How you can help ?

I am completely estatic , that you have thought of helping me. If you wish to do so, kindly find me and my kernels on KAGGLE, at [https://kaggle.com/justdvnsh](https://kaggle.com/justdvnsh)