Code Trans T5 Large Code Comment Generation Java Multitask Finetune

🚀 CodeTrans Model for Java Code Documentation Generation

This is a pre-trained model for the Java programming language, leveraging the T5 large model architecture. It was initially released in this repository. The model is trained on tokenized Java code functions and performs optimally with such inputs.

✨ Features

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

📦 Installation

No specific installation steps are provided in the original document.

💻 Usage Examples

Basic Usage

Here is 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_large_code_comment_generation_java_multitask_finetune"),
    tokenizer=AutoTokenizer.from_pretrained("SEBIS/code_trans_t5_large_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.

📚 Documentation

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. 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 V3-8 for 25,000 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 achieves the following results on different programming languages (in BLEU score):

Test results :

Created by Ahmed Elnaggar | LinkedIn and Wei Ding | LinkedIn