Aris Zaharia@ariscoi

We Fine-Tuned Qwen2.5-7b instruct for medical use. Methodology: Datasets: Fine-Tuning : qiaojin/PubMedQA (pqa_labeled) , LinhDuong/chatdoctor-200k (1000 samples) Evaluation : qiaojin/PubMedQA (pqa_labeled) (500 samples , LinhDuong/chatdoctor-200k (500 samples), GBaker/MedQA-USMLE-4-options (500 samples) Fine-tuned with LoRA : rank-8, Evaluation Criteria: We used Gemma 2 as a judge with the labels from the datasets as The Standard Refference. We evaluated Qwen base and our Model( Qwen Medical ) independently to see improvments. Each answer was then graded by the LLM Judge(Gemma 2) on this scale: Grading Rubric: 1: Completely incorrect/unsafe. 2: Major hallucinations. 3: Correct but lacks detail. 4: Highly accurate, minor flaws. 5: Medically flawless and perfectly structured. Qwen Base got a score of 2.04 while Qwen Medical got a score of 2.16 showing a 6% increase.

Grab your ideas, spin up compute, and start building what’s next! Start with $100 in AMD Developer Cloud credits and begin building right away. Get access to hands-on tutorials, workshops, and curated learning resources to help you build faster and go further. To unlock credits and join the hackathon, sign up for the AMD AI Developer Program. 🤝 Go solo or team up — ship something real. 💰 $21,500+ total prize pool & 1 × AMD Radeon™ AI PRO R9700 GPU up for grabs. 📍 On-site Venue (May 9-10): MindsDB SF AI Collective 3154 17th St, San Francisco, California, USA 📍On-site participation is by invitation only. Travel and accommodation expenses will not be covered. 🧑‍💻 Developers, builders, and innovators - bring your boldest ideas and let’s turn them into real AI systems on AMD.