st-taro's picture
Duplicate from st-taro/csen346_temp
5b7e9c7

A mixture of datasets focused on Source/Intermediate Representation training for multilingual code generation. For SCU CSEN346.

Usage: Pass to continued_pretrain.py with

  --dataset_name './tokenized_dataset'

Sources: UKPLab/SLTrans bigcode/the-stack allenai/peS2o open-web-math/open-web-math