Instructions to use danushkv/medgemma-cholec-inst with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use danushkv/medgemma-cholec-inst with PEFT:

from peft import PeftModel
from transformers import AutoModelForCausalLM

base_model = AutoModelForCausalLM.from_pretrained("google/medgemma-1.5-4b-it")
model = PeftModel.from_pretrained(base_model, "danushkv/medgemma-cholec-inst")

Transformers

How to use danushkv/medgemma-cholec-inst with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="danushkv/medgemma-cholec-inst")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("danushkv/medgemma-cholec-inst", dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use danushkv/medgemma-cholec-inst with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "danushkv/medgemma-cholec-inst"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "danushkv/medgemma-cholec-inst",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/danushkv/medgemma-cholec-inst

SGLang

How to use danushkv/medgemma-cholec-inst with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "danushkv/medgemma-cholec-inst" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "danushkv/medgemma-cholec-inst",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "danushkv/medgemma-cholec-inst" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "danushkv/medgemma-cholec-inst",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use danushkv/medgemma-cholec-inst with Docker Model Runner:
```
docker model run hf.co/danushkv/medgemma-cholec-inst
```

Model Card for Model ID

A fine-tuned version of MedGemma-1.5 on surgical tools of cholecystectomy surgical dataset

Model Details

Model Description

This is a fine-tuned Medgemma-1.5-4B-it model on image-text pairs of surgical images from the colecystectomy surgical procedure. The model can be prompted with an image with the text prompt "explain in detail the surgical scene inlcuding the position of the surgical tools". A short description of the scene is obtained as an output. These text prompts were used to further fine-tune image-text diffusion models for surgical context grounded image generation.

Uses

The model can be prompted with a surgical image and a corresponding text is obtained which describes the surgical tools present in the scene. Based on the set temperature, the information about anatomies can also be obtained. However, be cautious that certain hallucinations about the scene can be wrong

Training Details

For details on training this model, please refer to this repo

Downloads last month: 1

Model tree for danushkv/medgemma-cholec-inst

Base model

google/medgemma-1.5-4b-it

Adapter

(44)

this model

Collection including danushkv/medgemma-cholec-inst

SurgFasT

Collection

Fast tracking surgical science with MedGemma • 3 items • Updated Feb 24