tulas/gemma-3-270m-medical

Medical Fine-tuned Model

This model is a fine-tuned version of gemma-3-270m-it using LoRA (Low-Rank Adaptation) on medical data just for testing purpose

Model Details

• Base Model: google/gemma-3-270m
• Fine-tuning Method: LoRA (Low-Rank Adaptation)
• Domain: Medical/Healthcare
• Merged: Yes, LoRA adapters have been merged with the base model

Training Information

• Training Steps: 813
• Learning Rate: 3e-4
• LoRA Rank: 64
• LoRA Alpha: 16
• Target Modules: q_proj, k_proj, v_proj, o_proj

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("tulas/gemma-3-270m-medical")
tokenizer = AutoTokenizer.from_pretrained("tulas/gemma-3-270m-medical")

# Generate text
inputs = tokenizer("Patient presents with chest pain and", return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=100)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

Intended Use

This model is NOT intended for medical text generation but for testing purpose only

Limitations

• This model should not be used for actual medical diagnosis
• Always consult healthcare professionals for medical decisions
• Model outputs should be verified by medical experts

License

This model is released under the Apache 2.0 license.