TASTY stands for: A [T]ransformer based [A]pproach to [S]pace and [T]ime complexit[y].
This is a dataset of code snippets with their complexities, both space and time.
This is part of extended dataset that covers more than 10 programming languages.
You can find the original TASTY (TASTY v1) dataset here.
TASTY v1 was collected as part of our work on the paper called TASTY, published at the ICLR DL4Code workshop, a few years back.
We plan to… See the full description on the dataset page:
https://huggingface.co/datasets/Banana-Leopard/TASTY.plus.plus.