ajibawa-2023/Java-Code-Large

1---2license: mit3task_categories:4- text-generation5language:6- en7tags:8- code9- java10size_categories:11- 10M<n<100M12---13 14**Java-Code-Large**15 16Java-Code-Large is a large-scale corpus of publicly available Java source code comprising more than **15 million** java codes. The dataset is designed to support research in large language model (LLM) pretraining, code intelligence, software engineering automation, and program analysis.17 18By providing a high-volume, language-specific corpus, Java-Code-Large enables systematic experimentation in Java-focused model training, domain adaptation, and downstream code understanding tasks.19 20 21**1. Introduction**22 23Large-scale code corpora have become fundamental resources for training and evaluating machine learning models for code-related tasks. While multilingual code datasets exist, there is increasing interest in language-specialized corpora to:24 25- Improve domain-specific performance26 27- Reduce cross-language noise28 29- Enable controlled experimental settings30 31- Support Java-specific tooling and research32 33Java-Code-Large addresses this need by providing a dedicated Java-only dataset at substantial scale.34 35 36**2. Dataset Composition**37 38Programming Language: Java39 40File Count: 15M+ Java files41 42File Format: .jsonl43 44Content Types:45 46- Classes47 48- Interfaces49 50- Enums51 52- Methods53 54- Annotations55 56- JavaDoc comments57 58- Exception handling structures59 60- Generics and concurrency constructs61 62The dataset consists of source code extracted from publicly accessible open-source repositories.63 64 65**3. Intended Research Applications**66 67   683.1 Pretraining69 70- Training code foundation models from scratch71 72- Continued pretraining of existing LLMs73 74- Java-specialized language modeling75 76  77 783.2 Fine-Tuning and Adaptation79 80- Code completion systems81 82- Automated refactoring tools83 84- IDE copilots85 86- Java-specific conversational assistants87 88  89 903.3 Code Intelligence Tasks91 92- Code summarization93 94- Code-to-text generation95 96- Bug detection97 98- Vulnerability detection99 100- Clone detection101 102- Code similarity modeling103 104- Static and structural analysis105 106  107 1083.4 Software Engineering Research109 110- Empirical studies of Java programming patterns111 112- Tokenization and AST modeling experiments113 114 115 116 117Thanks to open source community for all the guidance & support!!118 119