SEBIS/code_trans_t5_base_source_code_summarization_python_transfer_learning_finetune

The CodeTrans model is based on the t5-base model architecture . It has its own SentencePiece vocabulary model . It used transfer-learning pre-training on 7 unsupervised datasets in the software development domain . It is then fine-tuned on the source code summarization task for python code snippets . The model could be used to generate the description for the python function . It can be used on unparsed and untokenized python code . It works best with tokenized python functions, but performance should be better if the python code is tokenized . It was trained on a single TPU Pod V3-8 for 500,000 steps in total, using sequence,