tomhu/RL4TG-Qwen2.5-3B-OPD-14B-Teacher
The tomhu/RL4TG-Qwen2.5-3B-OPD-14B-Teacher is a 3.1 billion parameter language model based on Qwen2.5-3B-Instruct, specifically fine-tuned for Java unit-test generation. It utilizes Online Policy Distillation (OPD) with Qwen2.5-Coder-14B-Instruct as the teacher model. This model is optimized for research in automated test generation, particularly for the Defects4J project, offering specialized capabilities in creating Java unit tests.
Loading preview...
tomhu/RL4TG-Qwen2.5-3B-OPD-14B-Teacher Overview
This model is a specialized 3.1 billion parameter language model, building upon the Qwen/Qwen2.5-3B-Instruct base. Its primary distinction lies in its fine-tuning for Java unit-test generation using an advanced technique called Online Policy Distillation (OPD).
Key Capabilities & Features
- Specialized Test Generation: Designed specifically to generate unit tests for Java codebases.
- Online Policy Distillation (OPD): Leverages
Qwen/Qwen2.5-Coder-14B-Instructas a powerful teacher model to guide its learning process, enhancing its test generation quality. - Defects4J Training: Trained on a project-disjoint split of the Defects4J dataset, indicating its focus on practical, bug-related test cases.
- Qwen2.5 Chat Template: Utilizes the standard Qwen2.5 chat prompt format for interaction.
Intended Use Cases
- Research in Automated Unit Test Generation: This model is explicitly intended for academic and research purposes related to automatically generating unit tests for Java applications.
- Code Quality & Reliability Studies: Can be used in experiments to explore the effectiveness of generated tests in identifying software defects.
It's important to note that while the model generates tests, these outputs require compilation and execution within the target project's build environment to be fully validated.