Paradigm-based Artificial Neural Dailogue Agent or PANDA

A LSTM AI Language Model trained on 25K Tokens of English Language and C-Progamming Language which can predict next sequence of words and it can also write some basic codes to some extent.

PANDA (Paradigm-based Artificial Neural Dialogue Agent) is an implementation of a Language Model (LM) using Keras and LSTM (Long Short-Term Memory) neural networks. This LM is designed for text generation based on a provided set of prompts.

It combines data preprocessing, neural network architecture, training, and text generation features to create a versatile LM capable of generating text based on input prompts. These features make it a valuable tool for various natural language processing tasks and creative text generation applications.

Usage

In Repository, there are two files located at Colab Notebook folder :

• PANDA.ipynb : This is the main experiment file which contains all the code for training and testing the model. You can run this file on Google Colab.
• Panda_Code_Refactored.ipynb : This file contains the code for main refactored code for the model. You can run this file on Google Colab.

Also, You can use this model by importing PANDA.py file in your project and use it as a class. Training this model on your own local machine takes a lot of time and resources. So, Therefore we have a saved model of it as PANDA.h5 you can find it at Saved Model/panda.h5 or if you want to train it by your self then I recommend you to use Google Colab for training this model.

Spinnet of code for using PANDA class in your project:

    from PANDA.transform import PANDA

    model = PANDA(['Train/Prompts/train_prompts.txt', 50])
    model.pretrained('saved_model/panda-25k-2.5-lstm-lm')

    for i in range(5): # inference for 5 turns
        user_prompt = input('==> ')
        print(model.infer(user_input=user_prompt))

Contributing

Contributions are welcome! For bug reports or requests please submit an issue. For new feature contribution please submit a pull request. If you would like to contribute to the project, here are some ways you can help:

• Improve the documentation
• Report bugs
• Fix bugs and submit pull requests
• Write, clarify, or fix documentation
• Suggest or add new features