autocomplete-ml

This repository contains a character-level Transformer notebook (mini-GPT) for Java autocomplete experiments.

Recommended environment setup

Pick one of the two approaches below.

1) Using a Python venv (pip)

# create and activate venv
python3 -m venv .venv
source .venv/bin/activate

# install dependencies
pip install -r requirements.txt

Note: PyTorch may require a specific wheel for CUDA or MPS. If the pip install -r requirements.txt command does not install a GPU-enabled build, follow the official PyTorch instructions for your platform: https://pytorch.org/get-started/locally/

2) Using conda

conda env create -f environment.yml
conda activate autocomplete-ml

For CUDA machines you might instead want to install PyTorch via conda with the correct pytorch-cuda package (see PyTorch docs). The provided environment.yml uses pip to install the same requirements.txt for simplicity.

Quick checks

• Open and run 01_char_level_transformer.ipynb.
• The notebook already contains a device-selection snippet that prints which device PyTorch is using (MPS/CUDA/CPU).

Notes

• requirements.txt contains minimal packages used in the notebook (PyTorch, numpy, jupyter). Add more packages if you extend the notebook.
• To pin exact versions for reproducibility, replace the >= pins with exact versions (e.g. torch==2.1.0).