Auto code is a NLP model that can predict python code snippets with the provided code to it. It is a LSTM (Long Short Term Memory model) with word Embedding.
It is trained with 606 rows of code snippets on a Macbook M1 Pro with a Total embedding word size of 237. The major thing I have tried in the project is to make the tokenizer of TensorFlow detect Special characters.
Special characters are converted to texts and then are tokenized and is reversed back to symbols after prediction.
This model can be trained for other languages just by creating code snippets of the respective language CSV file like dataset.csv file
python3 -m pip install TensorFlow
pip install pandas
pip install numpy
pip install matplotlib
- Tensorflow
- Pandas
- Numpy
- Matlab
To train the model with more data just add more code snippets in each cell of dataset.csv or add more data as python list in dataprocessing.ipynb file and run code to append data to csv file
- Train = .8
- Validiation = .2
- Accuracy
- Loss