SEBIS/code_trans_t5_small_commit_generation

The CodeTrans model is based on the t5-small model architecture . It has its own SentencePiece vocabulary model . It can be used on unparsed and untokenized commit changes . The model could be used to generate git commit message for the git commit changes or be fine-tuned on other relevant tasks . It works best with tokenized git commit . The code is trained on tokenized commit . It is based in-built on single-task training on Git Commit Message Generation dataset . For the test, different models achieves the following results on different programming languages (in BLEU score) For example, the code is generated using Transformers SummarizationPipeline,