spignelon / cleverterm

CleverTerm is a prototype LSTM model trained on bash history, capable of generating suggestions for the next Linux command based on previous user input. The project is licensed under GPL v3 and uses bash history data from various sources, including Kaggle datasets and GitHub repositories.

CleverTerm

CleverTerm is a prototype project that utilizes a LSTM model trained on bash history to suggest the next word in a command line interface. The project is unreleased and is currently in its alpha version.

Model Summary

Model: "sequential"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 embedding (Embedding)       (None, None, 14)          347858    
                                                                 
 lstm (LSTM)                 (None, None, 100)         46000     
                                                                 
 lstm_1 (LSTM)               (None, 100)               80400     
                                                                 
 dense (Dense)               (None, 100)               10100     
                                                                 
 dense_1 (Dense)             (None, 24847)             2509547   
                                                                 
=================================================================
Total params: 2,993,905
Trainable params: 2,993,905
Non-trainable params: 0
_________________________________________________________________                                                                 

The model was trained using TensorFlow and has a total of 2,993,905 parameters, all of which are trainable.

Usage

The trained model can be found at https://github.com/spignelon/CleverTerm/releases/download/v0.0.1-alpha/cleverterm_model.h5. More information about the project, including the code used to train the model, can be found in the following Jupyter notebook: https://github.com/spignelon/CleverTerm/blob/main/CleverTerm%20LSTM.ipynb.

Data Source

The bash history data used to train the model was taken from this Kaggle dataset, as well as various .bash_history files uploaded to GitHub and GitHub gists.

Future Development

At the moment, CleverTerm is an experimental project and its future development is uncertain. However, we may continue working on it in the future to improve its accuracy and functionality.

License

This project is licensed under the GNU General Public License v3.0

cleverterm's Issues

Couldn't get tensorflow working with GPU in Anaconda Navigator

Stopped the training of the model because on CPU training was very slow, but couldn't get the latest version of tensorflow to work with GPU, cudatoolkit installed successfully, but couldn't install nvidia-cudnn-cu11==8.6.0.163

Followed official docs, asked multiple people and searched the entire internet, but couldn't get it to work with GPU

https://www.tensorflow.org/install/pip#windows-wsl2

nvidia-cudnn-cu11==8.6.0.163 not found