Instructions to use CryptoGod97/meddroid-v9-lora with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use CryptoGod97/meddroid-v9-lora with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("unsloth/medgemma-4b-it-unsloth-bnb-4bit") model = PeftModel.from_pretrained(base_model, "CryptoGod97/meddroid-v9-lora") - Notebooks
- Google Colab
- Kaggle
MedDroid-v9 LoRA adapter
The LoRA adapter behind MedDroid-v9 (GGUF: CryptoGod97/meddroid-v9), published so MedDroid-v9 can be fine-tuned further (for example in the MedDroid-v9 hackathon).
- Base: Google's MedGemma-4B-it (Unsloth 4-bit copy), LoRA r=8 on the language layers, vision tower frozen
- Load with Unsloth:
FastModel.from_pretrained("CryptoGod97/meddroid-v9-lora", load_in_4bit=True)
Licence: built on MedGemma, so Google's Health AI Developer Foundations terms apply. MedDroid's additions are free for research and non-commercial use; commercial use needs MedDroid's permission (support@medicalandroid.com).
Educational use only. Not a medical device. Decision support for qualified doctors, not a diagnosis.
- Downloads last month
- 32
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐ Ask for provider support