CodeTrans model for source code summarization Python

Pretrained model on programming language python using the t5 large model architecture. It was first released in this repository. This model is trained on tokenized python code functions: it works best with tokenized python functions.

Model description

This CodeTrans model is based on the t5-large model. It has its own SentencePiece vocabulary model. It used multi-task training on 13 supervised tasks in the software development domain and 7 unsupervised datasets.

Intended uses & limitations

The model could be used to generate the description for the Python function or be fine-tuned on other Python code tasks. It can be used on unparsed and untokenized Python code. However, if the Python code is tokenized, the performance should be better.

How to use

Here is how to use this model to generate Python function documentation using Transformers SummarizationPipeline:

from transformers import AutoTokenizer, AutoModelWithLMHead, SummarizationPipeline

pipeline = SummarizationPipeline(
    model=AutoModelWithLMHead.from_pretrained("SEBIS/code_trans_t5_large_source_code_summarization_python_multitask"),
    tokenizer=AutoTokenizer.from_pretrained("SEBIS/code_trans_t5_large_source_code_summarization_python_multitask", skip_special_tokens=True),
    device=0
)

tokenized_code =  '''with open ( CODE_STRING , CODE_STRING ) as in_file : buf = in_file . readlines ( )  with open ( CODE_STRING , CODE_STRING ) as out_file : for line in buf :          if line ==   " ; Include this text   " :              line = line +   " Include below  "          out_file . write ( line ) '''
pipeline([tokenized_code])

Run this example in colab notebook.

Training data

The supervised training tasks datasets can be downloaded on Link

Training procedure

Multi-task Training

The model was trained on a single TPU Pod V3-8 for 80,000 steps, using sequence length 512 (batch size 4096). It has a total of approximately 220M parameters and was trained using the encoder-decoder architecture. The optimizer used is AdaFactor with inverse square root learning rate schedule for pre-training. (We have trained in total 260,000 steps.)

Evaluation results

For the code documentation tasks, different models achieves the following results on different programming languages (in BLEU score):

Test results :

Language / Model	Python	SQL	C#
CodeTrans-ST-Small	8.45	17.55	19.74
CodeTrans-ST-Base	9.12	15.00	18.65
CodeTrans-TF-Small	10.06	17.71	20.40
CodeTrans-TF-Base	10.94	17.66	21.12
CodeTrans-TF-Large	12.41	18.40	21.43
CodeTrans-MT-Small	13.11	19.15	22.39
CodeTrans-MT-Base	13.37	19.24	23.20
CodeTrans-MT-Large	13.24	19.40	23.57
CodeTrans-MT-TF-Small	12.10	18.25	22.03
CodeTrans-MT-TF-Base	10.64	16.91	21.40
CodeTrans-MT-TF-Large	12.14	19.98	21.10
State of the art	--	18.40	20.50

Created by Ahmed Elnaggar | LinkedIn and Wei Ding | LinkedIn