Show HN: Autocomplete Python Code with Transformers

3 pointsby vpjhace 6 años2 comments

This is a small project we created to train a character level autoregressive transformer (or LSTM) model to predict Python source code. We trained it on GitHub repositories found on awesome pytorch list.

Github repo: https://github.com/lab-ml/python_autocomplete

You can try training on Google Colab: https://colab.research.google.com/github/lab-ml/python_autocomplete/blob/master/notebooks/train.ipynb

Here are some sample evaluations/visualizations of the trained model: https://colab.research.google.com/github/lab-ml/python_autocomplete/blob/master/notebooks/evaluate.ipynb

Working on a simple VSCode extension to test this out. Will open source it soon on the same repository.

2 comments

vpjhace 6 años

Links:

Github repo: https://github.com/lab-ml/python_autocomplete

Training notebook: https://colab.research.google.com/github/lab-ml/python_autoc...

Evaluation notebook: https://colab.research.google.com/github/lab-ml/python_autoc...

alexanderrofailhace 6 años

https://www.kite.com/ does this too, claiming 47% less keystrokes. They are a good benchmark to set for seeing how efficient the model can get.