Haiintel/HaiJava-Surgeon-Qwen2.5-Coder-7B-SFT-v1

Haiintel/HaiJava-Surgeon-Qwen2.5-Coder-7B-SFT-v1 is a 7.6 billion parameter Supervised Fine-Tuned (SFT) model, based on Qwen/Qwen2.5-Coder-7B-Instruct, specifically optimized for Java bug-fixing. This model demonstrates a significant improvement in Java code correction, achieving an 82.28% pass@1 on the MultiPL-E Java benchmark, a 14.56% increase over its base model. It is designed for direct inference and as an improved baseline for further specialized fine-tuning in Java development tasks.