medgemma-brats-experiments

🧠 Medgemma-brats-experiments

This repository contains fine-tuning experiments using unsloth/medgemma-4b-it on BraTS and TextBraTS datasets for brain MRI and radiology text understanding. The experiments explore different LoRA configurations and their effects on domain adaptation, language specificity, and catastrophic forgetting.

---

🧩 Experimental Setup

• Base model: unsloth/medgemma-4b-it
• Datasets:
  • 🧬 Jupitern52/TextBraTS
  • 🧠 Kaggle BraTS20 Dataset
• Framework: Unsloth (LoRA fine-tuning)
• Training type: SFT (Supervised Fine-Tuning)
• Epochs: 1–2
• Loss curve:

---

⚙️ LoRA Configurations

---

📊 Quantitative Results

---

🧠 Qualitative Comparisons

Below are qualitative results comparing outputs across configurations using real generation excerpts.

🧩 Example 1 — Brain MRI Slice

Ground truth:

“Lesion in right frontal and parietal lobes with mixed signals. Edema and necrosis observed. Ventricular compression present.”

🖼️ Image:

---

🔬 Example 2 — Histology / Blood Sample

🖼️ Image:

---

🐶 Example 3 — Natural Image (Puppies)

🖼️ Image:

---

💬 Observations

• Fine-tuning on domain data improves medical accuracy, especially MRI lesion descriptions.
• Higher LoRA ranks (e.g., r=16) slightly improve quantitative metrics but cause catastrophic forgetting — the model loses general descriptive ability on unrelated domains.
• For small or specialized datasets, lighter LoRA configurations (r=1, α=4, 1–2 epochs) offer the best balance between specialization and generalization.
• In such limited-data scenarios, techniques like context engineering or prompt-based adaptation are often more efficient than full fine-tuning.
• This aligns with current research and discussions on when to fine-tune versus when to leverage better prompting strategies.

📚 References:

• Catastrophic Forgetting in LLMs: A Comparative Analysis Across Language Tasks — N. Haque, 2025
• When to Fine-Tune LLMs (and When Not To): A Practical Guide
• When Fine-Tuning Actually Makes Sense: A Developer’s Guide
• Prompt Engineering Guide — Context Engineering