Use requirements.txt to get all of the necessary libraries to run the model. You can use pip install -r requirements.txt to do so.
There is also the option of using an anaconda virtual environment. For this, you have to use conda commands with the file requirements_conda.yml as follows:
conda env create -f requirements_conda.ymland to activate the new environment:
conda activate NLP_tfaNOTE: In case you would like to perform the model training with GPU support on your machine, make sure that you do that on a Linux machine/emulation. The tensorflow package stopped GPU support for Windows machines after the release of version 2.11 and installing it by python -m pip install tensorflow<2.11 did not succeed at the moment of development of this project (apparently pip allows installing newer versions).
Everything we have currently can be defined in the following files:
main.py-- the part where the main control loop of the project takes place: from receiving the data set, to training the model and consequently testing it.architecture.py-- here we define the architectures of our Encoder and Decoder methods.training.py-- in this part, the our LSTM with attention architecture, including its functionality for training and code generation is defined.dataloader.py-- the code within this file is used to receive, pre-process (e.g., clean, tokenize and create emebeddings) our data set.evaluation_metrics.py-- this file is used for defining all the evalution metrics that will be used to assess the quality of the generated code.