Code Trans T5 Base Code Comment Generation Java Multitask Finetune

🚀 CodeTrans Model for Java Code Comment Generation

This is a pre - trained model for the Java programming language, which utilizes the T5 base model architecture. It was first introduced in this repository. This model is trained on tokenized Java code functions, and it performs optimally with such tokenized Java functions.

🚀 Quick Start

The model can be used to generate descriptions for Java functions or fine - tuned for other Java code tasks. It can handle unparsed and untokenized Java code, though performance is better with tokenized code.

Here's how to use this model to generate Java function documentation using Transformers SummarizationPipeline:

from transformers import AutoTokenizer, AutoModelWithLMHead, SummarizationPipeline

pipeline = SummarizationPipeline(
    model=AutoModelWithLMHead.from_pretrained("SEBIS/code_trans_t5_base_code_comment_generation_java_multitask_finetune"),
    tokenizer=AutoTokenizer.from_pretrained("SEBIS/code_trans_t5_base_code_comment_generation_java_multitask_finetune", skip_special_tokens=True),
    device=0
)

tokenized_code = "protected String renderUri ( URI uri ) { return uri . toASCIIString ( ) ; }"
pipeline([tokenized_code])

Run this example in colab notebook.

✨ Features

• Based on the t5 - base model with its own SentencePiece vocabulary model.
• Trained using multi - task training on 13 supervised tasks in the software development domain and 7 unsupervised datasets.
• Fine - tuned for the code comment generation task for Java functions/methods.

📚 Documentation

Model description

This CodeTrans model is based on the t5 - base 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. It is then fine - tuned on the code comment generation task for the Java function/method.

Intended uses & limitations

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

Training data

The supervised training tasks datasets can be downloaded on Link

Training procedure

Multi - task Pretraining

The model was trained on a single TPU Pod V3 - 8 for 260,000 steps in total, 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.

Fine - tuning

This model was then fine - tuned on a single TPU Pod V2 - 8 for 60,000 steps in total, using sequence length 512 (batch size 256), using only the dataset only containing Java code.

Evaluation results

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

Created by Ahmed Elnaggar | LinkedIn and Wei Ding | LinkedIn