Gemma 4 E2B — Empathetic Counseling Assistant
Fine-tuned Gemma 4 E2B for empathetic mental health counseling support.
Trained on MentalChat16K (16K counselor–client conversations covering 33 mental health topics).
Key Features
- Empathetic responses across anxiety, depression, grief, relationships, and more
- Bilingual — responds in English and Arabic (leveraging Gemma 4's native 140-language support)
- Lightweight — LoRA adapters (only ~0.5% of parameters trained)
- Deployable — GGUF export available for Ollama / llama.cpp
Training Details
- Base model: google/gemma-4-e2b-it
- Method: QLoRA (4-bit, rank 16, alpha 16)
- Dataset: ShenLab/MentalChat16K (16,113 Q&A pairs)
- Framework: Unsloth + TRL SFTTrainer
- Hardware: Google Colab (T4 16GB / A100 40GB)
Disclaimer
This model is for research and educational purposes only.
It is NOT a substitute for professional mental health care.
If you or someone you know is in crisis, please contact a licensed professional.
This gemma4 model was trained 2x faster with Unsloth
