File size: 308 Bytes
268e972 3848232 268e972 3848232 268e972 |
1 2 3 4 5 |
This model is designed for the Leo language and is based on `microsoft/codebert-base`.
With the code data written in the Leo language crawled from GitHub, the model is trained using self-supervised learning with triplet loss.
Anchor and positive code chunks are from the same code, while negatives are not. |